6 Objects, Collections, and OPAQUE Types

6.1.1 Introduction to Objects and Collections

The Oracle SQLJ implementation supports user-defined SQL object types, which are composite data structures, related SQL object reference types, and user-defined SQL collection types. Oracle objects and collections are composite data structures consisting of individual data elements.

The Oracle SQLJ implementation supports either strongly typed or weakly typed Java representations of object types, reference types, and collection types to use in iterators or host expressions. Strongly typed representations use a custom Java class that maps to a particular object type, reference type, or collection type and must implement either the Java Database Connectivity (JDBC) 2.0 standard java.sql.SQLData interface, for object types only, or the Oracle oracle.sql.ORAData interface. Either paradigm is supported by the Oracle Database 10g JPublisher utility, which you can use to automatically generate custom Java classes.

The term strongly typed is used where a particular Java type is associated with a particular SQL named type or user-defined type. For example, if there is a PERSON type, then a corresponding Person Java class will be associated with it.

Weakly typed representations use oracle.sql.STRUCT for objects, oracle.sql.REF for object references, or oracle.sql.ARRAY for collections. Alternatively, you can use standard java.sql.Struct, java.sql.Ref, or java.sql.Array objects in a weakly typed scenario.

The term weakly typed is used where a Java type is used in a generic way and can map to multiple SQL named types. The Java class or interface has no special information particular to any SQL type. This is the case for the oracle.sql.STRUCT, oracle.sql.REF, and oracle.sql.ARRAY types and the java.sql.Struct, java.sql.Ref, and java.sql.Array types.

Note that using Oracle extensions in your code requires the following:

• Use one of the Oracle JDBC drivers.

• Use default Oracle-specific code generation or, for ISO code generation, customize the profiles appropriately. For Oracle-specific generated code, no profiles are produced so customization is not applicable. Oracle JDBC application programming interfaces (APIs) are called directly through the generated Java code.

  
  

  Note: Oracle recommends the use of the default customizer, oracle.sqlj.runtime.util.OraCustomizer.

  
  

• Use the Oracle SQLJ run time when your application runs. The Oracle SQLJ run time and an Oracle JDBC driver are required whenever you use the Oracle customizer, even if you do not actually use Oracle extensions in your code.

For Oracle-specific semantics-checking, you must use an appropriate checker. The default checker, oracle.sqlj.checker.OracleChecker, acts as a front end and will run the appropriate checker based on your environment. This will be one of the Oracle-specific checkers if you are using an Oracle JDBC driver.

Custom Java Class Usage Notes

• This chapter primarily discusses the use of custom Java classes with user-defined types. However, classes implementing ORAData can be used for other Oracle SQL types as well. A class implementing ORAData can be used to perform any kind of desired processing or conversion in the course of transferring data between SQL and Java.

  
  

  See Also: "Additional Uses for ORAData Implementations"

  
  

• The SQLData interface is intended only for custom object classes. The ORAData interface can be used for any custom Java class.

Terminology Notes

• User-defined SQL object types and user-defined SQL collection types are referred to as user-defined types (UDTs).

• Custom Java classes for objects, references, and collections are referred to as custom object classes, custom reference classes, and custom collection classes, respectively.

6.1.2 Oracle Object Fundamentals

The Oracle SQL objects are composite data structures that group related data items, such as facts about each employee, into a single data unit. An object type is functionally similar to a Java class. You can populate and use any number of individual objects of a given object type, just as you can instantiate and use individual objects of a Java class.

For example, you can define an object type EMPLOYEE that has the attributes name of type CHAR, address of type CHAR, phonenumber of type CHAR, and employeenumber of type NUMBER.

Oracle objects can also have methods, or stored procedures, associated with the object type. These methods can be either static methods or instance methods and can be implemented either in PL/SQL or Java. Their signatures can include any number of input, output, or input-output parameters. All this depends on how they are initially defined

6.1.3 Oracle Collection Fundamentals

There are two categories of Oracle SQL collections:

• Variable-length arrays (VARRAY types)

• Nested tables (TABLE types)

Both categories are one-dimensional, although the elements can be complex object types. VARRAY types are used for one-dimensional arrays, and nested table types are used for single-column tables within an outer table. A variable of any VARRAY type can be referred to as a VARRAY. A variable of any nested table type can be referred to as a nested table.

A VARRAY, as with any array, is an ordered set of data elements, with each element having an index and all elements being of the same data type. The size of a VARRAY refers to the maximum number of elements. Oracle VARRAYs, as indicated by their name, are of variable size, but the maximum size of any particular VARRAY type must be specified when the VARRAY type is declared.

A nested table is an unordered set of elements. Nested table elements within a table can themselves be queried in SQL. A nested table, as with any table, is not created with any particular number of rows. This is determined dynamically.

6.1.4 Object and Collection Data Types

In Oracle Database 10g, user-defined object and collection definitions function as SQL data type definitions. You can use these data types, as with any other data type, in defining table columns, SQL object attributes, and stored procedure or function parameters. In addition, once you have defined an object type, the related object reference type can be used as any other SQL reference type.

For example, consider the EMPLOYEE Oracle object described in the preceding section. Once you have defined this object, it becomes an Oracle data type. You can have a table column of type EMPLOYEE just as you can have a table column of type NUMBER. Each row in an EMPLOYEE column contains a complete EMPLOYEE object. You can also have a column type of REF EMPLOYEE, consisting of references to EMPLOYEE objects.

Similarly, you can define a variable-length array MYVARR as VARRAY(10) of NUMBER and a nested table NTBL of CHAR(20). The MYVARR and NTBL collection types become Oracle data types, and you can have table columns of either type. Each row of a MYVARR column consists of an array of up to 10 numbers. Each row of an NTBL column consists of 20 characters.

6.2.1 Custom Java Class Interface Specifications

This section discusses specifications of the ORAData and ORADataFactory interfaces and the standard SQLData interface.

Oracle Database 10g includes a set of APIs for Oracle-specific custom Java class functionality for user-defined types: oracle.sql.ORAData and oracle.sql.ORADataFactory.

The oracle.sql.CustomDatum and oracle.sql.CustomDatumFactory interfaces used previously for this functionality are deprecated (as of Oracle9i), but still supported for backward compatibility. You must use the CustomDatum interfaces if you are working with an Oracle8i JDBC driver.

ORAData and ORADataFactory Specifications

Oracle provides the oracle.sql.ORAData interface and the related oracle.sql.ORADataFactory interface to use in mapping and converting Oracle object types, reference types, and collection types to custom Java classes.

Data is sent or retrieved in the form of an oracle.sql.Datum object, with the underlying data being in the format of the appropriate oracle.sql.Datum subclass, such as oracle.sql.STRUCT. This data is still in its SQL format. The oracle.sql.Datum object is just a wrapper.

The ORAData interface specifies a toDatum() method for data conversion from Java format to SQL format. This method takes as input your connection object and converts data to the appropriate oracle.sql.* representation. The connection object is necessary so that the JDBC driver can perform appropriate type checking and type conversions at run time. The ORAData and toDatum() specification is as follows:

interface oracle.sql.ORAData
{
   oracle.sql.Datum toDatum(java.sql.Connection c) throws SQLException;
}

The ORADataFactory interface specifies a create() method that constructs instances of your custom Java class, converting from SQL format to Java format. This method takes as input a Datum object containing the data and a type code, such as OracleTypes.RAW, indicating the SQL type of the underlying data. It returns an object of your custom Java class, which implements the ORAData interface. This object receives its data from the Datum object that was input. The ORADataFactory and create() specification is as follows:

interface oracle.sql.ORADataFactory
{
   oracle.sql.ORAData create(oracle.sql.Datum d, int sqlType) 
                      throws SQLException;
}

To complete the relationship between the ORAData and ORADataFactory interfaces, you must implement a static getORADataFactory() method in any custom Java class that implements the ORAData interface. This method returns an object that implements the ORADataFactory interface and that, therefore, can be used to create instances of your custom Java class. This returned object can itself be an instance of your custom Java class, and its create() method is used by the Oracle JDBC driver to produce further instances of your custom Java class, as necessary.

If you use JPublisher, then specifying -usertypes=oracle will result in JPublisher generating custom Java classes that implement the ORAData and ORADataFactory interfaces and the getORADataFactory() method. For backward compatibility, you have the option of using the JPublisher -compatible option in conjunction with -usertypes=oracle to use the CustomDatum and CustomDatumFactory interfaces instead.

ORAData Versus CustomDatum Interfaces

The oracle.jdbc interfaces were introduced in Oracle9i Database as replacements for the oracle.jdbc.driver classes. As a result, the oracle.sql.CustomDatum and oracle.sql.CustomDatumFactory interfaces, formerly used to access customized objects, are deprecated in favor of the oracle.sql.ORAData and oracle.sql.ORADataFactory interfaces. Like the CustomDatum interfaces, these can be used as an Oracle-specific alternative to the standard SQLData interface. The CustomDatum interfaces are still supported for backward compatibility.

CustomDatum and CustomDatumFactory have the following definitions:

public interface CustomDatum
{
  oracle.sql.Datum toDatum(
    oracle.jdbc.driver.OracleConnection conn
  ) throws SQLException;

public interface CustomDatumFactory
{
  oracle.sql.CustomDatum create(
    oracle.sql.Datum d, int sqlType
    ) throws SQLException;
}

The connection conn and type code sqlType are used as described for ORAData and ORADataFactory. Note, however, that CustomDatum uses the Oracle-specific OracleConnection type instead of the standard Connection type.

SQLData Specification

Standard JDBC 2.0 supplies the java.sql.SQLData interface to use in mapping and converting structured object types to Java classes. This interface is intended for mapping structured object types only, not object references, collections or arrays, or other SQL types.

The SQLData interface is a JDBC 2.0 standard, specifying a readSQL() method to read data into a Java object and a writeSQL() method to write to the database from a Java object. If you use JPublisher, then specifying -usertypes=jdbc will result in JPublisher generating custom Java classes that implement the SQLData interface.

For additional information about standard SQLData functionality, refer to the Sun Microsystems JDBC 2.0 or later API specification.

6.2.2 Custom Java Class Support for Object Methods

Methods of Oracle objects can be invoked from custom Java class wrappers. Whether the underlying stored procedure is written in PL/SQL or is written in Java and published to SQL is invisible to the user.

A Java wrapper method used to invoke a server method requires a connection to communicate with the server. The connection object can be provided as an explicit parameter or can be associated in some other way. For example, as an attribute of your custom Java class. If the connection object used by the wrapper method is a non-static attribute, then the wrapper method must be an instance method of the custom Java class in order to have access to the connection. Custom Java classes generated by JPublisher use this technique.

There are also issues regarding output and input-output parameters in methods of Oracle objects. If a stored procedure, that is, a SQL object method, modifies the internal state of one of its arguments, then the actual argument passed to the stored procedure is modified. In Java this is not possible. When a JDBC output parameter is returned from a stored procedure call, it must be stored in a newly created object. The original object identity is lost.

One way to return an output or input-output parameter to the caller is to pass the parameter as an element of an array. If the parameter is input-output, then the wrapper method takes the array element as input. After processing, the wrapper assigns the output to the array element. Custom Java classes generated by JPublisher use this technique, with each output or input-output parameter being passed in a one-element array.

When you use JPublisher, it implements wrapper methods by default. This is true for generated classes implementing either the SQLData interface or the ORAData interface. To disable this feature, set the JPublisher -methods flag to false.

6.2.3 Custom Java Class Requirements

Custom Java classes must satisfy certain requirements to be recognized by the Oracle SQLJ translator as valid host variable types and to enable type-checking by the translator.

Oracle Requirements for Classes Implementing ORAData

Oracle requirements for ORAData implementations are primarily the same for any kind of custom Java class, but vary slightly depending on whether the class is for mapping to objects, object references, collections, or some other SQL type.

These requirements are as follows:

• The class implements the oracle.sql.ORAData interface.

• The class implements the getORADataFactory() method that returns an oracle.sql.ORADataFactory object. The method signature is as follows:

  public static oracle.sql.ORADataFactory getORADataFactory();
  
  

  If using the deprecated CustomDatum interface, then the class implements the getFactory() method that returns an oracle.sql.CustomDatumFactory object. The method signature is as follows:

  public static oracle.sql.CustomDatumFactory getFactory();
  
  

• The class has a String constant, _SQL_TYPECODE, initialized to the oracle.jdbc.OracleTypes type code of the Datum subclass instance that toDatum() returns. The type code is:

  • For custom object classes:

    public static final int _SQL_TYPECODE = OracleTypes.STRUCT;
    
    

  • For custom reference classes:

    public static final int _SQL_TYPECODE = OracleTypes.REF;
    
    

  • For custom collection classes:

    public static final int _SQL_TYPECODE = OracleTypes.ARRAY;
    
    

  For other uses, some other type code might be appropriate. For example, for using a custom Java class to serialize and deserialize Java objects into or out of RAW fields, a _SQL_TYPECODE of OracleTypes.RAW is used.

  
  

  Note: The OracleTypes class simply defines a type code, which is an integer constant, for each Oracle data type. For standard SQL types, the OracleTypes entry is identical to the entry in the standard java.sql.Types type definitions class.

  
  
  
  

  See Also: "Serialized Java Objects"

  
  

• For custom Java classes with _SQL_TYPECODE of STRUCT, REF, or ARRAY, that is, for custom Java classes that represent objects, object references, or collections, the class has a constant that indicates the relevant user-defined type name. This is as follows:

  • Custom object classes and custom collection classes must have a String constant, _SQL_NAME, initialized to the SQL name you declared for the user-defined type, as follows:

    public static final String _SQL_NAME = UDT name;
    
    

    For example, the custom object class for a user-defined PERSON object will have the constant:

    public static final String _SQL_NAME = "PERSON";
    
    

    The same can be specified along with the schema, if appropriate, as follows:

    public static final String _SQL_NAME = "SCOTT.PERSON";
    
    

    The custom collection class for a collection of PERSON objects, which you have declared as PERSON_ARRAY, will have the constant:

    public static final String _SQL_NAME = "PERSON_ARRAY";
    
    

  • Custom reference classes must have a String constant, _SQL_BASETYPE, initialized to the SQL name you declared for the user-defined type being referenced, as follows:

    public static final String _SQL_BASETYPE = UDT name;
    
    

    The custom reference class for PERSON references will have the constant:

    public static final String _SQL_BASETYPE = "PERSON";
    
    

    For other ORAData uses, specifying a UDT name is not applicable.

Keep in mind the following usage notes:

• A collection type name reflects the collection type, not the base type. For example, if you have declared a VARRAY or nested table type, PERSON_ARRAY, for PERSON objects, then the name of the collection type that you specify for the _SQL_NAME entry is PERSON_ARRAY, not PERSON.

• When specifying the SQL type in a _SQL_NAME field, if the SQL type was declared in a case-sensitive way (in quotes), then you must specify the SQL name exactly as it was declared, such as CaseSensitive or SCOTT.CaseSensitive. Note that this differs from usage in a JPublisher input file, where the case-sensitive name must also appear in quotes. If you did not declare the SQL type in a case-sensitive way, that is, without no quotes, then you must specify the SQL name in all uppercase, such as ADDRESS or SCOTT.ADDRESS.

  JPublisher automatically generates the value of this field appropriately, according to case-sensitivity and the JPublisher -omit_schema_names setting, if applicable.

Requirements for Classes Implementing SQLData

The ISO SQLJ standard outlines requirements for type map definitions for classes implementing the SQLData interface. Alternatively, SQLData wrapper classes can identify associated SQL object types through the public static final fields.

Be aware of the following important points:

• Whether you use a type map or use alternative (non-standard) public static final fields to specify mappings, you must be consistent in your approach. Either use a type map that specifies all relevant mappings so that you do not require the public static final fields, or do not use a type map at all and specify all mappings through the public static final fields.

• SQLData, unlike ORAData, is for mapping structured object types only. It is not for object references, collections or arrays, or any other SQL types. If you are not using ORAData, then your only choices for mapping object references and collections are the weak java.sql.Ref and java.sql.Array types, respectively, or oracle.sql.REF and oracle.sql.ARRAY.

• SQLData implementations require a Java Development Kit (JDK) 1.2.x or later environment.

• When specifying the mapping from a SQL type to a Java type, if the SQL type was declared in a case-sensitive way, then you must specify the SQL name exactly as it was declared, such as CaseSensitive or SCOTT.CaseSensitive. Note that this differs from usage in a JPublisher input file, where the case-sensitive name must also appear in quotes. If you did not declare the SQL type in a case-sensitive way, then you must specify the SQL name in all uppercase, such as ADDRESS or SCOTT.ADDRESS.

Mapping Specified in Type Map Resource

First, consider the mapping representation according to the ISO SQLJ standard. Assume that Address, pack.Person, and pack.Manager.InnerPM, where InnerPM is an inner class of Manager, are three wrapper classes that implement java.sql.SQLData.

Then, you need to consider the following:

• You must use these classes only in statements that use explicit connection context instances of a declared connection context type. For example, assuming that this type is called SDContext:

  Address               a =...;
  pack.Person           p =...;
  pack.Manager.InnerPM pm =...;
  SDContext ctx = new SDContext(url,user,pwd,false);
  #sql [ctx] { ... :a ... :p ... :pm ... };
  
  

• The connection context type must have been declared using the with attribute typeMap that specifies an associated class implementing java.util.PropertyResourceBundle. In the preceding example, SDContext may be declared as follows:

  #sql public static context SDContext with (typeMap="SDMap");
  
  

• The type map resource must provide the mapping from SQL object types to corresponding Java classes that implement the java.sql.SQLData interface. This mapping is specified with entries of the following form:

  class.java_class_name=STRUCT sql_type_name
  
  

  The STRUCT keyword can also be omitted. In the example, the SDMap.properties resource file may contain the following entries:

  class.Address=STRUCT SCOTT.ADDRESS
  class.pack.Person=PERSON
  class.pack.Manager$InnerPM=STRUCT PRODUCT_MANAGER
  
  

  Although the period (.) separates package and class name, you must use the dollar sign ($) to separate an inner class name.

#sql public static iterator SDIter with (typeMap="SDMap");
...
SDContext sdctx = ...
SDIter sditer;
#sql [sdctx] sditer = { SELECT ...};

This mechanism of specifying mappings in a type map resource is more complicated than the non-standard alternative. Also, it is not possible to associate a type map resource with the default connection context. The advantage is that all the mapping information is placed in a single location, the type map resource. This means that the type mapping in an already compiled application can be easily adjusted at a later time, for example, to accommodate new SQL types and Java wrappers in an expanding SQL-Java type hierarchy.

Be aware of the following:

• You must employ the SQLJ runtime12 or runtime12ee library to use this feature. Type maps are represented as java.util.Map objects. These are exposed in the SQLJ run time API and, therefore, cannot be supported by the generic run time library.

• You must use the Oracle SQLJ run time and Oracle-specific code generation or profile customization if your SQLData wrapper classes occur as OUT or INOUT parameters in SQLJ statements. This is because the SQL type of such parameters is required for registerOutParameter() by the Oracle JDBC driver. Also, for OUT parameter type registration, the SQL type is "frozen in" by the type map in effect during translation.

• The SQLJ type map is independent of any JDBC type map you may be using on the underlying connection. Thus, you must be careful when you are mixing SQLJ and JDBC code if both use SQLData wrappers. However, you can easily extract the type map in effect on a given SQLJ connection context:

  ctx.getTypeMap();
  

Mapping Specified in Static Field of Wrapper Class

A class that implements SQLData can satisfy the following non-standard requirement:

• The Java class declares the String constant _SQL_NAME, which defines the name of the SQL type that is being wrapped by the Java class. In the example, the Address class would have the following field declaration:

  public static final String _SQL_NAME="SCOTT.ADDRESS";
  
  

  The following declaration would be in pack.Person:

  public static final String _SQL_NAME="PERSON";
  
  

  And the pack.Manager.InnerPM class would have the following:

  public static final String _SQL_NAME="PRODUCT_MANAGER";
  
  

Note that JPublisher always generates SQLData wrapper classes with the _SQL_NAME field. However, this field is ignored in SQLJ statements that reference a type map.

6.2.4 Compiling Custom Java Classes

You can include any .java files for your custom Java classes, whether ORAData or SQLData implementations, on the SQLJ command line together with the .sqlj files for your application. However, this is not necessary if the SQLJ -checksource flag is set to true, which is the default, and your classpath includes the directory where the custom Java source is located.

For example, if ObjectDemo.sqlj uses the ADDRESS and PERSON Oracle object types and you have produced custom Java classes for these objects, then you can run SQLJ as follows.

• If -checksource=true and the classpath includes the custom Java source location:

  % sqlj ObjectDemo.sqlj
  
  

• If -checksource=false (this is a single wraparound line):

  % sqlj ObjectDemo.sqlj Address.java AddressRef.java Person.java PersonRef.java
  
  

You also have the choice of using your Java compiler to compile custom .java source files directly. If you do this, then you must do it prior to translating .sqlj files.

6.2.5 Reading and Writing Custom Data

Through the use of custom Java class instances, the Oracle SQLJ and JDBC implementations allow you to read and write user-defined types as though they are built-in types. Exactly how this is accomplished is transparent to the user.

For the mechanics of how data is read and written, for both ORAData implementations and SQLData implementations, refer to the Oracle Database JDBC Developer's Guide and Reference.

6.2.6 Additional Uses for ORAData Implementations

To this point, discussion of custom Java classes has been for use as one of the following.

• Wrappers for SQL objects: custom object classes, for use with oracle.sql.STRUCT instances

• Wrappers for SQL references: custom reference classes, for use with oracle.sql.REF instances

• Wrappers for SQL collections: custom collection classes, for use with oracle.sql.ARRAY instances

It might be useful, however, to provide custom Java classes to wrap other oracle.sql.* types as well, for customized conversions or processing. You can accomplish this with classes that implement ORAData, but not SQLData, as in the following examples:

• Perform encryption and decryption or validation of data.

• Perform logging of values that have been read or are being written.

• Parse character columns, such as character fields containing URL information, into smaller components.

• Map character strings into numeric constants.

• Map data into more desirable Java formats, such as mapping a DATE field to java.util.Date format.

• Customize data representation, for example, data in a table column is in feet, but you want it represented in meters after it is selected.

• Serialize and deserialize Java objects, for example, into or out of RAW fields.

General Use of ORAData: BetterDate.java

This example shows a class that implements the ORAData interface to provide a customized representation of Java dates and can be used instead of java.sql.Date.

import java.util.Date;
import oracle.sql.ORAData;
import oracle.sql.DATE;
import oracle.sql.ORADataFactory;
import oracle.jdbc.OracleTypes;

// a Date class customized for user's preferences:
//      - months are numbers 1..12, not 0..11
//      - years are referred to through four-digit numbers, not two.

public class BetterDate extends java.util.Date
             implements ORAData, ORADataFactory {
  public static final int _SQL_TYPECODE = OracleTypes.DATE;
  
  String[]monthNames={"JAN", "FEB", "MAR", "APR", "MAY", "JUN",
                      "JUL", "AUG", "SEP", "OCT", "NOV", "DEC"};
  String[]toDigit={"0", "1", "2", "3", "4", "5", "6", "7", "8", "9"};

  static final BetterDate _BetterDateFactory = new BetterDate();

  public static ORADataFactory getORADataFactory() { return _BetterDateFactory;}

  // the current time...
  public BetterDate() {
    super();
  }

  public oracle.sql.Datum toDatum(java.sql.Connection conn) {
    return new DATE(toSQLDate());
  }

  public oracle.sql.ORAData create(oracle.sql.Datum dat, int intx) {
    if (dat==null) return null;
    DATE DAT = ((DATE)dat);
    java.sql.Date jsd = DAT.dateValue();
    return new BetterDate(jsd);
  }
   
  public java.sql.Date toSQLDate() {
    java.sql.Date retval;
    retval = new java.sql.Date(this.getYear()-1900, this.getMonth()-1,
             this.getDate());
    return retval;
  }
  public BetterDate(java.sql.Date d) {
    this(d.getYear()+1900, d.getMonth()+1, d.getDate());
  }
  private static int [] deconstructString(String s) {
    int [] retval = new int[3];
    int y,m,d; char temp; int offset;
    StringBuffer sb = new StringBuffer(s);
    temp=sb.charAt(1);
    // figure the day of month
    if (temp < '0' || temp > '9') {
      m = sb.charAt(0)-'0';
      offset=2;
    } else {
      m = (sb.charAt(0)-'0')*10 + (temp-'0');
      offset=3;
    }

    // figure the month
    temp = sb.charAt(offset+1);
    if (temp < '0' || temp > '9') {
      d = sb.charAt(offset)-'0';
      offset+=2;
    } else {
      d = (sb.charAt(offset)-'0')*10 + (temp-'0');
      offset+=3;
    }

    // figure the year, which is either in the format "yy" or "yyyy"
    // (the former assumes the current century)
    if (sb.length() <= (offset+2)) {
      y = (((new BetterDate()).getYear())/100)*100 +
          (sb.charAt(offset)- '0') * 10 +
          (sb.charAt(offset+1)- '0');
    } else {
      y = (sb.charAt(offset)- '0') * 1000 +
          (sb.charAt(offset+1)- '0') * 100 +
          (sb.charAt(offset+2)- '0') * 10 +
          (sb.charAt(offset+3)- '0');
    }
    retval[0]=y;
    retval[1]=m;
    retval[2]=d;
//    System.out.println("Constructing date from string as: "+d+"/"+m+"/"+y);
    return retval;
  }
  private BetterDate(int [] stuff) {
    this(stuff[0], stuff[1], stuff[2]);
  }
  // takes a string in the format: "mm-dd-yyyy" or "mm/dd/yyyy" or
  // "mm-dd-yy" or "mm/dd/yy" (which assumes the current century)
  public BetterDate(String s) {
    this(BetterDate.deconstructString(s));
  }

  // years are as '1990', months from 1..12 (unlike java.util.Date!), date
  // as '1' to '31' 
  public BetterDate(int year, int months, int date) {
    super(year-1900,months-1,date);
  }
  // returns "Date: dd-mon-yyyy"
  public String toString() { 
    int yr = getYear();
    return getDate()+"-"+monthNames[getMonth()-1]+"-"+
      toDigit[(yr/1000)%10] + 
      toDigit[(yr/100)%10] + 
      toDigit[(yr/10)%10] + 
      toDigit[yr%10];
//    return "Date: " + getDate() + "-"+getMonth()+"-"+(getYear()%100);
  }
  public BetterDate addDays(int i) {
    if (i==0) return this;
    return new BetterDate(getYear(), getMonth(), getDate()+i);
  }
  public BetterDate addMonths(int i) {
    if (i==0) return this;
    int yr=getYear();
    int mon=getMonth()+i;
    int dat=getDate();
    while(mon<1) { 
      --yr;mon+=12;
    }
    return new BetterDate(yr, mon,dat);
  }
  // returns year as in 1996, 2007
  public int getYear() {
    return super.getYear()+1900;
  }
  // returns month as 1..12
  public int getMonth() {
    return super.getMonth()+1;
  }
  public boolean equals(BetterDate sd) {
    return (sd.getDate() == this.getDate() &&
            sd.getMonth() == this.getMonth() &&
            sd.getYear() == this.getYear());
  }
  // subtract the two dates; return the answer in whole years
  // uses the average length of a year, which is 365 days plus
  // a leap year every 4, except 100, except 400 years =
  // = 365 97/400 = 365.2425 days = 31,556,952 seconds
  public double minusInYears(BetterDate sd) {
    // the year (as defined in the preceding text) in milliseconds
    long yearInMillis = 31556952L;
    long diff = myUTC()-sd.myUTC();
    return (((double)diff/(double)yearInMillis)/1000.0);
  }
  public long myUTC() {
    return Date.UTC(getYear()-1900, getMonth()-1, getDate(),0,0,0);
  }
  
  // returns <0 if this is earlier than sd
  // returns = if this == sd
  // else returns >0
  public int compare(BetterDate sd) {
    if (getYear()!=sd.getYear()) {return getYear()-sd.getYear();}
    if (getMonth()!=sd.getMonth()) {return getMonth()-sd.getMonth();}
    return getDate()-sd.getDate();
  }
}

6.4.1 What JPublisher Produces

When you use JPublisher to generate custom Java classes, you can use either an ORAData implementation, for custom object classes, custom reference classes, or custom collection classes, or a SQLData implementation, for custom object classes only. An ORAData implementation will also implement the ORADataFactory interface, for creating instances of the custom Java class.

This is controlled by how you set the JPublisher -usertypes option. A setting of -usertypes=oracle specifies an ORAData implementation, and a setting of -usertypes=jdbc specifies a SQLData implementation.

ORAData Implementation

When you run JPublisher for a user-defined object type and use the ORAData implementation for your custom object class, JPublisher automatically creates the following:

• A custom object class, typically in a .sqlj source file, to act as a type definition to correspond to your Oracle object type

  This class includes accessor methods for each attribute. For example, getFoo() and setFoo() are the accessor methods for the attribute foo. In addition, JPublisher by default will generate wrapper methods in your class that invoke the associated Oracle object methods executing in the server. However, this can be disabled by setting -methods=false. In this case, JPublisher produces no wrapper methods and generates .java files instead of .sqlj files for custom objects.

• A related custom reference class for object references to your Oracle object type

  This class includes a getValue() method that returns an instance of your custom object class and a setValue() method that updates an object value in the database, taking as input an instance of the custom object class.

  A strongly typed reference class is always generated, regardless of whether the SQL object type uses references.

• Custom classes for any object or collection attributes of the top-level object

  This is necessary so that attributes can be materialized in Java whenever an instance of the top-level class is materialized.

When you run JPublisher for a user-defined collection type, choosing the ORAData implementation, JPublisher automatically creates the following:

• A custom collection class to act as a type definition to correspond to your Oracle collection type

  This class includes overloaded getArray() and setArray() methods to retrieve or update a collection as a whole, a getElement() method and setElement() method to retrieve or update individual elements of a collection, and additional utility methods.

• A custom object class for the elements, if the elements of the collection are objects

  This is necessary so that object elements can be materialized in Java whenever an instance of the collection is materialized.

JPublisher-generated custom Java classes in any of these categories implement the ORAData interface, the ORADataFactory interface, and the getORADataFactory() method.

Strongly Typed Object References for ORAData Implementations

For Oracle ORAData implementations, JPublisher always generates strongly typed object reference classes as opposed to using the weakly typed oracle.sql.REF class. This is to provide greater type safety and to mirror the behavior in SQL, where object references are strongly typed. The strongly typed classes, with names like PersonRef for references to PERSON objects, are essentially wrappers for the REF class.

In these strongly typed REF wrappers, there is a getValue() method that produces an instance of the SQL object that is referenced, in the form of an instance of the corresponding Java class. (Or, in the case of inheritance, perhaps as an instance of a subclass of the corresponding Java class.) For example, if there is a PERSON SQL object type with a corresponding Person Java class, then there will also be a PersonRef Java class. The getValue() method of the PersonRef class would return a Person instance containing the data for a PERSON object in the database.

Whenever a SQL object type has an attribute that is an object reference, the Java class corresponding to the object type would have an attribute that is an instance of a Java class corresponding to the appropriate reference type. For example, if there is a PERSON object with a MANAGER REF attribute, then the corresponding Person Java class will have a ManagerRef attribute.

SQLData Implementation

When you run JPublisher for a user-defined object type and choose the SQLData implementation for your custom object class, JPublisher will produce a custom object class to act as a type definition to correspond to your Oracle object type. This class will include the following:

• Accessor methods for each attribute

• Implementations of the readSQL() and writeSQL() methods of the standard SQLData interface

• Wrapper methods that invoke the Oracle object methods executing in the server, unless you specify -methods=false when you run JPublisher

Because the SQLData interface is intended only for objects, however, and not for references or collections, JPublisher will not generate a custom reference class for references to the Oracle object type. You will have to use standard weakly typed java.sql.Ref instances or perhaps oracle.sql.REF instances, if you do not require portability. Note that REF instances, like custom reference class instances, have Oracle extension methods, getValue() and setValue(), to read or write instances of the referenced object. Standard Ref instances do not have this functionality.

Similarly, because you cannot use a SQLData implementation for a custom collection class, you must use standard weakly typed java.sql.Array instances or perhaps oracle.sql.ARRAY instances, if you do not require portability. Array and ARRAY instances, like custom collection class instances, have getArray() functionality to read the collection as a whole or in part, but do not have the element-level access and writability offered by the getElement() and setElement() methods of the custom collection class.

6.4.2 Generating Custom Java Classes

This section discusses key JPublisher command-line functionality for specifying the user-defined types that you want to map to Java and for specifying object class names, collection class names, attribute type mappings, and wrapper methods. These key points can be summarized as follows:

• Specify the implementation to use, through the JPublisher -usertypes option.

• Specify user-defined types to map to Java. You can specify the custom object and custom collection class names for JPublisher to use, or you can accept the default names. Use the JPublisher -sql, -user, and -case options, as appropriate.

• Optionally specify attribute type mappings through the JPublisher -xxxtypes options: -numbertypes, -builtintypes, and -lobtypes.

• Choose whether or not JPublisher will create wrapper methods, in particular for Oracle object methods. Use the JPublisher -methods flag, which is enabled by default.

Choose the Implementation for Generated Classes

Before running JPublisher, consider whether you want the generated classes to implement the Oracle ORAData interface or the standard SQLData interface. Using SQLData will likely make your code more portable, but using ORAData offers a number of advantages, including no need for type maps.

Remember the following:

• You must use ORAData implementations for custom collection classes. The SQLData interface does not support collections.

• Strongly typed reference classes are always generated for ORAData custom object class implementations, but not for SQLData custom object class implementations. The SQLData interface does not support strongly typed object references. Use the weak java.sql.Ref type or oracle.sql.REF type instead.

Use the JPublisher -usertypes option to specify which interface you want your classes to implement. A setting of -usertypes=oracle, which is the default, specifies the ORAData interface, while a setting of -usertypes=jdbc specifies the SQLData interface.

The following JPublisher command-line examples will result in implementation of ORAData, CustomDatum, and SQLData, respectively (assume % is a system prompt).

% jpub -usertypes=oracle ... <other option settings>

% jpub -usertypes=oracle -compatible=customdatum ... <other option settings>

% jpub -usertypes=jdbc ... <other option settings>

JPublisher will ignore a -compatible=customdatum or -compatible=oradata setting if -usertypes=jdbc.

Specify User-Defined Types to Map to Java

In using JPublisher to create custom Java classes, use the -sql option to specify the user-defined SQL types that you want to map to Java. You can either specify the custom object class names and custom collection class names, or you can accept the defaults.

The default names of your top-level custom classes, the classes that will correspond to the user-defined type names you specify to the -sql option, are identical to the user-defined type names as you enter them on the JPublisher command line. Because SQL names in the database are not case-sensitive by default, you can capitalize them to ensure that your class names are capitalized according to Java convention. For example, if you want to generate a custom class for employee objects, you can run JPublisher as follows:

% jpub -sql=Employee ...

The default names of other classes, such as for the home_address objects that are attributes of employee objects, are determined by the JPublisher -case option. If you do not set the -case option, then it is set to mixed. This means that the default for the custom class name is to capitalize the initial character of the corresponding user-defined type name and the initial character of every word unit thereafter. JPublisher interprets underscores (_), dollar signs ($), and any characters that are illegal in Java identifiers as word-unit separators. These characters are discarded in the process.

For example, for Oracle object type home_address, JPublisher would create class HomeAddress in a HomeAddress.sqlj or .java source file.

On the JPublisher command line, use the following syntax for the -sql option:

-sql=udt1<:mapclass1><,udt2<:mapclass2>>,...,<udtN<:mapclassN>> ...

Note that you can specify multiple actions in a single option setting.

Use the -user option to specify the database schema. Following is an example:

% jpub -sql=Myobj,mycoll:MyCollClass -user=scott/tiger

For the MYOBJ Oracle object, this command will name it as you typed it, creating Myobj.sqlj source to define a Myobj class. For the MYCOLL Oracle collection, this command will create source MyCollClass.java to define a MyCollClass class.

You can optionally specify schema names in the -sql option, such as in the following example that specifies the scott schema:

% jpub -sql=scott.Myobj,scott.mycoll:MyCollClass -user=scott/tiger

You cannot specify custom reference class names. JPublisher automatically derives them by adding Ref to custom object class names. This is relevant to ORAData implementations only. For example, if JPublisher produces the Myobj.sqlj Java source to define the Myobj custom object class, then it will also produce the MyobjRef.java Java source to define a MyobjRef custom reference class.

To create custom Java classes for the object and collection types defined in "User-Defined Types", you can run JPublisher as follows:

% jpub -user=scott/tiger -sql=Address,Person,Phone_array,Participant_t,Module_t,Moduletbl_t

Alternatively, to explicitly specify custom object class and custom collection class names, run it as follows:

% jpub -user=scott/tiger -sql=Address,Person,phone_array:PhoneArray,
participant_t:ParticipantT,module_t:ModuleT,moduletbl_t:ModuletblT

Note that each of the preceding two examples is a single wraparound command line.

The second example will produce the following Java source files: Address.sqlj, AddressRef.java, Person.sqlj, PersonRef.java, PhoneArray.java, ParticipantT.sqlj, ParticipantTRef.java, ModuleT.sqlj, ModuleTRef.java, and ModuletblT.java. Examples of some of these source files are provided in "JPublisher Custom Java Class Examples".

So that it knows how to populate the custom Java classes, JPublisher connects to the specified schema to determine attributes of your specified object types or elements of your specified collection types.

If you want to change how JPublisher uses character case in default names for the methods and attributes that it generates, including lower-level custom Java class names for attributes that are objects or collections, then you can accomplish this using the -case option. There are four possible settings:

• -case=mixed (default)

  The following will be uppercase: the first character of every word unit of a class name, every word unit of an attribute name, and every word unit after the first word unit of a method name. All other characters are in lowercase. JPublisher interprets underscores (_), dollar signs ($), and any characters that are illegal in Java identifiers as word-unit separators. These characters are discarded in the process.

• -case=same

  Character case is unchanged from its representation in the database. Underscores and dollar signs are retained, and illegal characters are discarded.

• -case=upper

  Lowercase letters are converted to uppercase. Underscores and dollar signs are retained, and illegal characters are discarded.

• -case=lower

  Uppercase letters are converted to lowercase. Underscores and dollar signs are retained, and illegal characters are discarded.

Specify Type Mappings

JPublisher offers several choices for how to map user-defined types and their attribute and element types between SQL and Java.

JPublisher categorizes SQL types into the following groups, with corresponding JPublisher options as noted:

• Numeric types: Anything stored as SQL type NUMBER

  Use the JPublisher -numbertypes option to specify type-mapping for numeric types.

• Large object (LOB) types: SQL types BLOB and CLOB

  Use the JPublisher -lobtypes option to specify type-mapping for LOB types.

• Built-in types: Anything stored as a SQL type not covered by the preceding categories, for example, CHAR, VARCHAR2, LONG, and RAW

  Use the JPublisher -builtintypes option to specify type-mapping for built-in types.

JPublisher defines the following type-mapping modes:

• JDBC mapping (setting jdbc): Uses standard default mappings between SQL types and Java native types. This setting is valid for the -numbertypes, -lobtypes, and -builtintypes options.

• Oracle mapping (setting oracle): Uses corresponding oracle.sql types to map to SQL types. This setting is valid for the -numbertypes, -lobtypes, and -builtintypes options.

• Object-JDBC mapping (setting objectjdbc): This is an extension of JDBC mapping. Where relevant, object-JDBC mapping uses numeric object types from the standard java.lang package, such as java.lang.Integer, Float, and Double, instead of primitive Java types, such as int, float, and double. The java.lang types are nullable, but the primitive types are not. This setting is valid for the -numbertypes option only.

• BigDecimal mapping (setting bigdecimal): Uses java.math.BigDecimal to map to all numeric attributes. This is appropriate if you are dealing with large numbers, but do not want to map to the oracle.sql.NUMBER type. This setting is valid for the -numbertypes option only.

If you do not specify mappings for the attribute types of a SQL object type or the element types of a SQL collection type, then JPublisher uses the following defaults:

• For numeric types, object-JDBC mapping is the default mapping.

• For LOB types, Oracle mapping is the default mapping.

• For built-in type types, JDBC mapping is the default mapping.

If you want alternate mappings, then use the -numbertypes, -lobtypes, and -builtintypes options as necessary, depending on the attribute types you have and the mappings you desire.

If an attribute type is itself a SQL object type, then it will be mapped according to the -usertypes setting.

Table 6-1 summarizes JPublisher categories for SQL types, the mapping settings relevant for each category, and the default settings.

Generate Wrapper Methods

In creating custom object classes to map Oracle objects to Java, the -methods option instructs JPublisher whether to include Java wrappers for Oracle object methods. The default -methods=true setting generates wrappers and also results in JPublisher generating a .sqlj file instead of a .java file for a custom object class, unless the underlying SQL object actually has no methods.

Wrapper methods generated by JPublisher are always instance methods, even when the original object methods are static. The following example shows how to set the -methods option:

% jpub -sql=Myobj,mycoll:MyCollClass -user=scott/tiger -methods=true

This will use default naming. The Java method names will be derived in the same fashion as custom Java class names, except that the initial character will be lowercase. For example, by default an object method name of CALC_SAL results in a Java wrapper method of calcSal(). Alternatively, you can specify desired Java method names, but this requires use of a JPublisher INPUT file.

Regarding Overloaded Methods

If you run JPublisher for an Oracle object that has an overloaded method where multiple signatures have the same corresponding Java signature, then JPublisher will generate a uniquely named method for each signature. It accomplishes this by appending _n to function names, where n is a number. This is to ensure that no two methods in the generated custom Java class have the same name and signature. For example, consider the SQL functions defined in creating a MY_TYPE object type:

CREATE OR REPLACE TYPE my_type AS OBJECT
(
   ...

   MEMBER FUNCTION myfunc(x INTEGER)
      RETURN my_return IS
      BEGIN
         ...
      END;

   MEMBER FUNCTION myfunc(y SMALLINT)
      RETURN my_return IS
      BEGIN
         ...
      END;
   ...
);

Without precaution, both definitions of myfunc result in the following name and signature in Java:

myfunc(Integer)

This is because both INTEGER and SMALLINT in SQL map to the Java Integer type.

Instead, JPublisher might call one myfunc_1 and the other myfunc_2. The _n is unique for each. In simple cases it will likely be _1, _2, and so on, but it might sometimes be arbitrary, other than being unique for each.

Generate Custom Java Classes and Map Alternate Classes

You can use JPublisher to generate a custom Java class but instruct it to map the object type or collection type to an alternative class instead of to the generated class.

A typical scenario is to treat JPublisher-generated classes as superclasses, extend them to add functionality, and map the object types to the subclasses. For example, presume you have an Oracle object type ADDRESS and want to produce a custom Java class for it that has functionality beyond what is produced by JPublisher. You can use JPublisher to generate a JAddress custom Java class for extending it to produce a MyAddress class. Under this scenario you will add any special functionality to MyAddress and will want JPublisher to map ADDRESS objects to that class, not to the JAddress class. You will also want JPublisher to produce a reference class for MyAddress, not JAddress.

JPublisher has functionality to streamline the process of mapping to alternative classes. Use the following syntax in your -sql option setting:

-sql=object_type:generated_class:map_class

For the preceding example, use this setting:

-sql=ADDRESS:JAddress:MyAddress

This generates class JAddress in source file JAddress.sqlj (or possibly .java) but does the following:

• Maps the ADDRESS object type to the MyAddress class, not to the JAddress class. Therefore, if you retrieve an object from the database that has an ADDRESS attribute, then this attribute will be created as an instance of MyAddress in Java. Or, if you retrieve an ADDRESS object directly, then you will retrieve it into a MyAddress instance.

• Creates a MyAddressRef class in MyAddressRef.java, instead of creating a JAddressRef class.

• Creates an initial version of the MyAddress class in a MyAddress.sqlj source file (or possibly MyAddress.java), unless the file already exists, in which case it is not changed.

MyAddress subclasses JAddress. In order to implement the extended functionality for MyAddress, you can start with the JPublisher-generated MyAddress source file, editing it as desired.

For further discussion about extending JPublisher-generated classes (continuing the preceding example), refer to "Extending Classes Generated by JPublisher".

6.4.3 JPublisher INPUT Files and Properties Files

JPublisher supports the use of special INPUT files and standard properties files to specify type mappings and additional option settings.

Using JPublisher INPUT Files

You can use the JPublisher -input command-line option to specify an INPUT file for JPublisher to use for additional type mappings. SQL in an INPUT file is equivalent to -sql on the command line, and the AS or GENERATE...AS syntax is equivalent to the command-line colon syntax. Use the following syntax, specifying just one mapping per SQL command:

SQL udt1 <GENERATE GeneratedClass1> <AS MapClass1>
SQL udt2 <GENERATE GeneratedClass2> <AS MapClass2>
...

This generates GeneratedClass1 and GeneratedClass2, but maps udt1 to MapClass1 and udt2 to MapClass2.

SQL "CaseSenstiveType" AS CaseSensitiveType

SQL SCOTT."CaseSensitiveType" AS CaseSensitiveType

SQL "Scott"."CaseSensitiveType" AS CaseSensitiveType

INPUT File Example

In the following example, JPublisher will pick up the -user option from the command line and go to INPUT file myinput.in for type mappings.

Command line:

% jpub -input=myinput.in -user=scott/tiger

Contents of INPUT file myinput.in:

SQL Myobj
SQL mycoll AS MyCollClass
SQL employee GENERATE Employee AS MyEmployee

This accomplishes the following:

• User-defined type MYOBJ gets the custom object class name Myobj because that is how you typed it. JPublisher creates source Myobj.sqlj (or possibly Myobj.java, if Myobj has no methods) and MyobjRef.java.

• User-defined type MYCOLL is mapped to MyCollClass. JPublisher creates a MyCollClass.java source file.

• User-defined type EMPLOYEE is mapped to the MyEmployee class. JPublisher creates source Employee.sqlj (or possibly Employee.java) and MyEmployeeRef.java, as well as an initial version of MyEmployee.sqlj (or .java) unless the file already exists. If you retrieve an object from the database that has an EMPLOYEE attribute, then this attribute would be created as an instance of MyEmployee in Java. Or, if you retrieve an EMPLOYEE object directly, presumably you will retrieve it into a MyEmployee instance. You are responsible for the MyEmployee code, but for convenience you can start with the JPublisher-generated MyEmployee source file and edit it to implement your specialized functionality for EMPLOYEE objects in Java. MyEmployee subclasses the Employee class.

Using JPublisher Properties Files

You can use the JPublisher -props command-line option to specify a properties file for JPublisher to use for additional type mappings and other option settings.

In a properties file, jpub. (including the period) is equivalent to the command-line "-" (single-dash), and other syntax remains the same. Specify only one option per line.

For type mappings, for example, jpub.sql is equivalent to -sql. You can specify multiple mappings in a single jpub.sql setting. Alternatively, you can use multiple jpub.sql options. The effect would be cumulative, as for multiple -sql options on the command line.

Properties File Example

In the following example, JPublisher will pick up the -user option from the command line and go to the jpub.properties properties file for type mappings and the attribute-mapping option.

Command line:

% jpub -props=jpub.properties -user=scott/tiger

Contents of properties file jpub.properties:

jpub.sql=Myobj,mycoll:MyCollClass,employee:Employee:MyEmployee
jpub.usertypes=oracle

This produces the same results as the preceding input-file example, explicitly specifying the oracle mapping setting.

6.4.4 Creating Custom Java Classes and Specifying Member Names

In generating custom Java classes, you can specify the names of any attributes or methods of the custom class. However, this cannot be specified on the JPublisher command line. You must specify it in a JPublisher INPUT file using TRANSLATE syntax, as follows:

SQL udt <GENERATE GeneratedClass> <AS MapClass> <TRANSLATE membername1 AS Javaname1> <, membername2 AS Javaname2> ...

TRANSLATE pairs (membernameN AS JavanameN) are separated by commas.

For example, presume the EMPLOYEE Oracle object type has an ADDRESS attribute that you want to call HomeAddress, and a GIVE_RAISE method that you want to call giveRaise(). Also presume that you want to generate an Employee class but map EMPLOYEE objects to a MyEmployee class that you will create. (This is not related to specifying member names, but provides a full example of INPUT file syntax.)

SQL employee GENERATE Employee AS MyEmployee 
TRANSLATE address AS HomeAddress, GIVE_RAISE AS giveRaise

6.4.5 JPublisher Implementation of Wrapper Methods

This section describes how JPublisher generates wrapper methods and how wrapper method calls are processed at run time.

Generation of Wrapper Methods

The following points describe how JPublisher generates wrapper methods:

• JPublisher-generated wrapper methods are implemented in SQLJ. Therefore, whenever -methods=true, the custom object class will be defined in a .sqlj file instead of in a .java file, assuming the object type defines methods. Run SQLJ to translate the .sqlj file.

• All wrapper methods generated by JPublisher are implemented as instance methods. This is because a database connection is required for you to invoke the corresponding server method. Each instance of a JPublisher-generated custom Java class has a connection associated with it.

Run Time Execution of Wrapper Method Calls

The following points describe what JPublisher-generated Java wrapper methods execute at run time. In this discussion, "Java wrapper method" refers to a method in the custom Java object, while "wrapped SQL method" refers to the SQL object method that is wrapped by the Java wrapper method.

• The custom Java object is converted to a SQL object and passed to the database, where the wrapped SQL method is invoked. After this method invocation, the new value of the SQL object is returned to Java in a new custom Java object, either as a function return from the wrapped SQL method, if the SQL method is a stored procedure, or as an array element in an additional output parameter, if the SQL method is a stored function and there already is a function return.

• Any output or input-output parameter is passed as the element of a one-element array. If the parameter is input-output, then the wrapper method takes the array element as input. After processing, the wrapper assigns the output to the array element.

6.4.6 JPublisher Custom Java Class Examples

This section provides examples of JPublisher-generated ORAData implementations for the following user-defined types:

• A custom object class (Address, corresponding to the Oracle object type ADDRESS) and related custom reference class (AddressRef)

• A custom collection class (ModuletblT, corresponding to the Oracle collection type MODULETBL_T)

Assume that the -methods option has its default true setting and that the ADDRESS type has methods, so that a .sqlj file is generated for the Address class.

Custom Object Class: Address.sqlj

Following is an example of the source code that JPublisher generates for a custom object class. Implementation details have been omitted.

In this example, unlike in "Creating Object Types", assume the Oracle object ADDRESS has only the street and zip_code attributes.

package bar;

import java.sql.SQLException;
import java.sql.Connection;
import oracle.jdbc.OracleTypes;
import oracle.sql.ORAData;
import oracle.sql.ORADataFactory;
import oracle.sql.Datum;
import oracle.sql.STRUCT;
import oracle.jpub.MutableStruct;

public class Address implements ORAData, ORADataFactory
{
  public static final String _SQL_NAME = "SCOTT.ADDRESS";
  public static final int _SQL_TYPECODE = OracleTypes.STRUCT;

  public static ORADataFactory getORADataFactory()
  { ... }

  /* constructors */
  public Address()
  { ... }

  public Address(String street, java.math.BigDecimal zip_code)
                throws SQLException
  { ... }

  /* ORAData interface */
  public Datum toDatum(Connection c) throws SQLException
  { ... }

  /* ORADataFactory interface */
  public ORAData create(Datum d, int sqlType) throws SQLException
  { ... }

  /* accessor methods */
  public String getStreet() throws SQLException
  { ... }

  public void setStreet(String street) throws SQLException
  { ... }


  public java.math.BigDecimal getZipCode() throws SQLException
  { ... }

  public void setZipCode(java.math.BigDecimal zip_code) throws SQLException
  { ... }

}

Custom Reference Class: AddressRef.java

Following is an example of the source code that JPublisher generates for a custom reference class to be used for references to ADDRESS objects. Implementation details have been omitted.

package bar;

import java.sql.SQLException;
import java.sql.Connection;
import oracle.jdbc.OracleTypes;
import oracle.sql.ORAData;
import oracle.sql.ORADataFactory;
import oracle.sql.Datum;
import oracle.sql.REF;
import oracle.sql.STRUCT;

public class AddressRef implements ORAData, ORADataFactory
{
  public static final String _SQL_BASETYPE = "SCOTT.ADDRESS";
  public static final int _SQL_TYPECODE = OracleTypes.REF;

  public static ORADataFactory getORADataFactory()
  { ... }

  /* constructors */
  public AddressRef()
  { ... }

  public static AddressRef(ORAData o) throws SQLException
  { ... }

  /* ORAData interface */
  public Datum toDatum(Connection c) throws SQLException
  { ... }

  /* ORADataFactory interface */
  public ORAData create(Datum d, int sqlType) throws SQLException
  { ... }

  public static AddressRef cast(ORAData o) throws SQLException
  { ... }

  public Address getValue() throws SQLException
  { ... }

  public void setValue(Address c) throws SQLException
  { ... }
}

Custom Collection Class: ModuletblT.java

Following is an example of the source code that JPublisher generates for a custom collection class. Implementation details have been omitted.

import java.sql.SQLException;
import java.sql.Connection;
import oracle.jdbc.OracleTypes;
import oracle.sql.ORAData;
import oracle.sql.ORADataFactory;
import oracle.sql.Datum;
import oracle.sql.ARRAY;
import oracle.sql.ArrayDescriptor;
import oracle.jpub.runtime.MutableArray;

public class ModuletblT implements ORAData, ORADataFactory
{
  public static final String _SQL_NAME = "SCOTT.MODULETBL_T";
  public static final int _SQL_TYPECODE = OracleTypes.ARRAY;

  public static ORADataFactory getORADataFactory()
  { ... }

  /* constructors */
  public ModuletblT()
  { ... }

  public ModuletblT(ModuleT[] a)
  { ... }

  /* ORAData interface */
  public Datum toDatum(Connection c) throws SQLException
  { ... }

  /* ORADataFactory interface */
  public ORAData create(Datum d, int sqlType) throws SQLException
  { ... }

  public String getBaseTypeName() throws SQLException
  { ... }

  public int getBaseType() throws SQLException
  { ... }

  public ArrayDescriptor getDescriptor() throws SQLException
  { ... }

  /* array accessor methods */
  public ModuleT[] getArray() throws SQLException
  { ... }

  public void setArray(ModuleT[] a) throws SQLException
  { ... }

  public ModuleT[] getArray(long index, int count) throws SQLException
  { ... }

  public void setArray(ModuleT[] a, long index) throws SQLException
  { ... }

  public ModuleT getObjectElement(long index) throws SQLException
  { ... }

  public void setElement(ModuleT a, long index) throws SQLException
  { ... }
}

6.4.7 Extending Classes Generated by JPublisher

You might want to enhance the functionality of a custom Java class generated by JPublisher by adding methods and transient fields. You can accomplish this by extending the JPublisher-generated class.

For example, suppose you want JPublisher to generate the JAddress class from the ADDRESS SQL object type. You also want to use a MyAddress class to represent ADDRESS objects and implement special functionality. The MyAddress class must extend JAddress.

Another way to enhance the functionality of a JPublisher-generated class is to simply add methods to it. However, adding methods to the generated class is not recommended if you anticipate running JPublisher at some future time to regenerate the class. If you run JPublisher to regenerate a class that you have modified in this way, then you would have to save a copy and manually merge your changes back in.

JPublisher Functionality for Extending Generated Classes

The syntax to have JPublisher generate JAddress but map to MyAddress is as follows:

-sql=ADDRESS:JAddress:MyAddress

Or, use the following in an INPUT file:

SQL ADDRESS GENERATE JAddress AS MyAddress

As a result of this, JPublisher will generate the MyAddressRef reference class, rather than JAddressRef, in MyAddressRef.java.

In addition, JPublisher alters the code it generates to implement the following functionality:

• The MyAddress class, instead of the JAddress class, is used to represent attributes whose SQL type is ADDRESS.

• The MyAddress class, instead of the JAddress class, is used to represent method arguments and function results whose type is ADDRESS.

• The MyAddress factory, instead of the JAddress factory, is used to construct Java objects whose SQL type is ADDRESS.

You would presumably use MyAddress similarly in any additional code that you write.

At run time, the Oracle JDBC driver will map any occurrences of ADDRESS data in the database to MyAddress instances, instead of to JAddress instances.

Requirements of Extended Classes

By default, JPublisher will create an initial version of the MyAddress user subclass in MyAddress.sqlj, if the original class uses methods and you are publishing these methods, or MyAddress.java, unless the file to be created already exists, in which case it will not be changed. You can edit this file as necessary to add your desired functionality.

MyAddress must have a no-argument constructor. The easiest way to construct a properly initialized object is to invoke the constructor of the superclass, either explicitly or implicitly.

As a result of extending the JPublisher-generated class, the subclass will inherit definitions of the _SQL_NAME field, which it requires, and the _SQL_TYPECODE field.

In addition, one of the following will be true.

• If the JPublisher-generated class implements the ORAData and ORADataFactory interfaces, then the subclass will inherit this implementation and the necessary toDatum() and create() functionality of the generated class. The subclass implements a getORADataFactory() method that returns an instance of your map class, such as a MyAddress object.

• If the JPublisher-generated class implements the SQLData interface, then the subclass will inherit this implementation and the necessary readSQL() and writeSQL() functionality of the generated class.

JPublisher-Generated Custom Object Class: JAddress.sqlj

The code for the JPublisher-generated JAddress class, implementing ORAData and ORADataFactory, is mostly identical to the code shown previously for the Address class, with the exception that mentions of Address are replaced by mentions of JAddress.

JPublisher-Generated Alternate Reference Class: MyAddressRef.java

Continuing the example in the preceding sections, consider code for the JPublisher-generated reference class, MyAddressRef (as opposed to JAddressRef, because MyAddress is the class that ADDRESS objects map to). This class also implements ORAData and ORADataFactory. The implementation is nearly identical to that of AddressRef.java, except for the change in class name and the fact that accessor methods use MyAddress instances instead of Address instances.

Extended Custom Object Class: MyAddress.sqlj

Again continuing the example, here is sample code for a MyAddress class that subclasses the JPublisher-generated JAddress class. The comments in the code show what is inherited from JAddress. Implementation details have been omitted.

import java.sql.SQLException;
import oracle.sql.ORAData;
import oracle.sql.ORADataFactory;
import oracle.sql.Datum;
import oracle.sql.STRUCT;
import oracle.jpub.runtime.MutableStruct;

public class MyAddress extends JAddress
{
  /* _SQL_NAME inherited from MyAddress */
  /* _SQL_TYPECODE inherited from MyAddress */

  static _myAddressFactory = new MyAddress();

  public static ORADataFactory getORADataFactory()
  {
    return _myAddressFactory;
  }

  /* constructor */
  public MyAddress()
  { super(); }

  /* ORAData interface */
  /* toDatum() inherited from JAddress */

  /* ORADataFactory interface */
  public ORAData create(oracle.sql.Datum d, int sqlType) throws SQLException
  { ... }

  /* accessor methods inherited from JAddress */

  /* Additional methods go here.  These additional methods (not shown)
     are the reason that JAddress was extended.
  */
}

6.5.1 Selecting Objects and Object References into Iterator Columns

This example uses a custom Java class and a custom reference class as iterator column types. Presume the following definition of the ADDRESS Oracle object type:

CREATE TYPE ADDRESS AS OBJECT
(  street VARCHAR(40),
   zip NUMBER );

And the following definition of the EMPADDRS table, which includes an ADDRESS column and an ADDRESS reference column:

CREATE TABLE empaddrs
(  name VARCHAR(60),
   home ADDRESS,
   loc REF ADDRESS );

Once you use JPublisher or otherwise create a custom Java class, Address, and custom reference class, AddressRef, corresponding to the ADDRESS Oracle object type, you can use Address and AddressRef in a named iterator as follows:

#sql iterator EmpIter (String name, Address home, AddressRef loc);

...
EmpIter ecur;
#sql ecur = { SELECT name, home, loc FROM empaddrs };
while (ecur.next()) {
   Address homeAddr = ecur.home();
   // Print out the home address.
   System.out.println ("Name: " + ecur.name() + "\n" +
                       "Home address: " + homeAddr.getStreet() + "   " +
                       homeAddr.getZip());
   // Now update the loc address zip code through the address reference.
   AddressRef homeRef = ecur.loc();
   Address location = homeRef.getValue();
   location.setZip(new BigDecimal(98765));
   homeRef.setValue(location);
   }
...

The ecur.home() method call extracts an Address object from the home column of the iterator and assigns it to the homeAddr local variable (for efficiency). The attributes of that object can then be accessed using standard Java dot syntax:

homeAddr.getStreet()

Use the getValue() and setValue() methods, standard with any JPublisher-generated custom reference class, to manipulate the location address (in this case its zip code).

6.5.2 Updating an Object

This example declares and sets an input host variable of the Address Java type to update an ADDRESS object in a column of the employees table. Both before and after the update, the address is selected into an output host variable of the Address type and printed for verification.

...
// Updating an object 

static void updateObject() 
{

   Address addr;
   Address new_addr;
   int empnum = 1001;

   try {
      #sql {
         SELECT office_addr
         INTO :addr
         FROM employees
         WHERE empnumber = :empnum };
      System.out.println("Current office address of employee 1001:");

      printAddressDetails(addr);

      /* Now update the street of address */

      String street ="100 Oracle Parkway";
      addr.setStreet(street);

      /* Put updated object back into the database */

      try {
         #sql {
            UPDATE employees
            SET office_addr = :addr
            WHERE empnumber = :empnum };
         System.out.println
            ("Updated employee 1001 to new address at Oracle Parkway.");

         /* Select new address to verify update */
      
         try {
            #sql {
               SELECT office_addr
               INTO :new_addr
               FROM employees
               WHERE empnumber = :empnum };
      
            System.out.println("New office address of employee 1001:");
            printAddressDetails(new_addr);

         } catch (SQLException exn) {
         System.out.println("Verification SELECT failed with "+exn); }
      
      } catch (SQLException exn) {
      System.out.println("UPDATE failed with "+exn); }

   } catch (SQLException exn) {
   System.out.println("SELECT failed with "+exn); }
}
...

Note the use of the setStreet() accessor method of the Address object. Remember that JPublisher provides such accessor methods for all attributes in any custom Java class that it produces.

This example uses the printAddressDetails() utility. The source code for this method is as follows:

static void printAddressDetails(Address a) throws SQLException
{

  if (a == null)  {
    System.out.println("No Address available.");
    return;
   }

   String street = ((a.getStreet()==null) ? "NULL street" : a.getStreet()) ;
   String city = (a.getCity()==null) ? "NULL city" : a.getCity();
   String state = (a.getState()==null) ? "NULL state" : a.getState();
   String zip_code = (a.getZipCode()==null) ? "NULL zip" : a.getZipCode();

   System.out.println("Street: '" + street + "'");
   System.out.println("City:   '" + city   + "'");
   System.out.println("State:  '" + state  + "'");
   System.out.println("Zip:    '" + zip_code + "'" );
}

6.5.3 Inserting an Object Created from Individual Object Attributes

This example declares and sets input host variables corresponding to attributes of PERSON and nested ADDRESS objects, then uses these values to insert a new PERSON object into the persons table in the database.

...
// Inserting an object

static void insertObject() 
{
   String new_name   = "NEW PERSON";
   int    new_ssn    = 987654;
   String new_street = "NEW STREET";
   String new_city   = "NEW CITY";
   String new_state  = "NS";
   String new_zip    = "NZIP";
  /*
   * Insert a new PERSON object into the persons table
   */
   try {
      #sql {
         INSERT INTO persons
         VALUES (PERSON(:new_name, :new_ssn,
         ADDRESS(:new_street, :new_city, :new_state, :new_zip))) };

      System.out.println("Inserted PERSON object NEW PERSON."); 

   } catch (SQLException exn) { System.out.println("INSERT failed with "+exn); }
}
...

6.5.4 Updating an Object Reference

This example selects a PERSON reference from the persons table and uses it to update a PERSON reference in the employees table. It uses simple input host variables to check attribute value criteria. The newly updated reference is then used in selecting the PERSON object to which it refers, so that information can be output to the user to verify the change.

...
// Updating a REF to an object

static void updateRef()
{
   int empnum = 1001;
   String new_manager = "NEW PERSON";

   System.out.println("Updating manager REF.");
   try {
      #sql {
         UPDATE employees
         SET manager = 
            (SELECT REF(p) FROM persons p WHERE p.name = :new_manager)
         WHERE empnumber = :empnum };

      System.out.println("Updated manager of employee 1001. Selecting back");

   } catch (SQLException exn) {
   System.out.println("UPDATE REF failed with "+exn); }

   /* Select manager back to verify the update */
   Person manager;

   try { 
      #sql {
         SELECT deref(manager)
         INTO :manager
         FROM employees e
         WHERE empnumber = :empnum };

      System.out.println("Current manager of "+empnum+":");
      printPersonDetails(manager);

   } catch (SQLException exn) {
   System.out.println("SELECT REF failed with "+exn); }

}
...

6.6.1 Accessing Nested Tables: TABLE syntax and CURSOR syntax

The Oracle SQLJ implementation supports the use of nested iterators to access data in nested tables. Use the CURSOR keyword in the outer SELECT statement to encapsulate the inner SELECT statement. This is shown in "Selecting Data from a Nested Table Using a Nested Iterator".

Oracle also supports use of the TABLE keyword to manipulate the individual rows of a nested table. This keyword informs Oracle that the column value returned by a subquery is a nested table, as opposed to a scalar value. You must prefix the TABLE keyword to a subquery that returns a single column value or an expression that yields a nested table.

The following example shows the use of the TABLE syntax:

UPDATE TABLE(SELECT a.modules FROM projects a WHERE a.id=555) b
       SET module_owner= 
       (SELECT ref(p) FROM employees p WHERE p.ename= 'Smith') 
       WHERE b.module_name = 'Zebra';

When you see TABLE used as it is here, realize that it is referring to a single nested table that has been selected from a column of an outer table.

6.6.2 Inserting a Row that Includes a Nested Table

This example shows an operation that manipulates the master level (outer table) and detail level (nested tables) simultaneously and explicitly. This inserts a row in the projects table, where each row includes a nested table of the MODULETBL_T type, which contains rows of MODULE_T objects.

First, the scalar values are set (id, name, start_date, duration), then the nested table values are set. This involves an extra level of abstraction, because the nested table elements are objects with multiple attributes. In setting the nested table values, each attribute value must be set for each MODULE_T object in the nested table. Finally, the owner values, initially set to null, are set in a separate statement.

// Insert Nested table details along with master details 

  public static void insertProject2(int id)  throws Exception 
  {
    System.out.println("Inserting Project with Nested Table details..");
    try {
      #sql { INSERT INTO Projects(id,name,owner,start_date,duration, modules) 
             VALUES ( 600, 'Ruby', null, '10-MAY-98',  300, 
             moduletbl_t(module_t(6001, 'Setup ', null, '01-JAN-98', 100),
                        module_t(6002, 'BenchMark', null, '05-FEB-98',20) ,
                        module_t(6003, 'Purchase', null, '15-MAR-98', 50),
                        module_t(6004, 'Install', null, '15-MAR-98',44),
                        module_t(6005, 'Launch', null,'12-MAY-98',34))) };
    } catch ( Exception e) {
      System.out.println("Error:insertProject2");
      e.printStackTrace();
    }

    // Assign project owner to this project 

    try {
      #sql { UPDATE Projects pr
          SET owner=(SELECT ref(pa) FROM participants pa WHERE pa.empno = 7698)
         WHERE pr.id=600 };
    } catch ( Exception e) {
      System.out.println("Error:insertProject2:update");
      e.printStackTrace();
    }
  }

6.6.3 Selecting a Nested Table into a Host Expression

This example presents an operation that works directly at the detail level of the nested table. Recall that ModuletblT is a JPublisher-generated custom collection class (ORAData implementation) for MODULETBL_T nested tables, ModuleT is a JPublisher-generated custom object class for MODULE_T objects, and MODULETBL_T nested tables contain MODULE_T objects.

A nested table of MODULE_T objects is selected from the modules column of the projects table into a ModuletblT host variable.

Following that, the ModuletblT variable (containing the nested table) is passed to a method that accesses its elements through its getArray() method, writing the data to a ModuleT[] array. All custom collection classes generated by JPublisher include a getArray() method. Then each element is copied from the ModuleT[] array into a ModuleT object, and individual attributes are retrieved through accessor methods (getModuleName(), for example) and then printed. All JPublisher-generated custom object classes include such accessor methods.

static ModuletblT mymodules=null;
  ...

  public static void getModules2(int projId)
  throws Exception 
  {
    System.out.println("Display modules for project " + projId );

    try {
      #sql {SELECT modules INTO :mymodules 
                           FROM projects  WHERE id=:projId };
      showArray(mymodules);
    } catch(Exception e) {
      System.out.println("Error:getModules2");
      e.printStackTrace();
    }
  }

  public static void showArray(ModuletblT a) 
  {
    try {
      if ( a == null )
        System.out.println( "The array is null" );
      else {
        System.out.println( "printing ModuleTable array object of size "
                             +a.length());
        ModuleT[] modules = a.getArray();

        for (int i=0;i<modules.length; i++) {
          ModuleT module = modules[i];
          System.out.println("module "+module.getModuleId()+
                ", "+module.getModuleName()+
                ", "+module.getModuleStartDate()+
                ", "+module.getModuleDuration());
        }
      }
    }
    catch( Exception e ) {
      System.out.println("Show Array");
      e.printStackTrace();
    }
  }

6.6.4 Manipulating a Nested Table Using TABLE Syntax

This example uses TABLE syntax to work at the detail level to access and update nested table elements directly, based on master-level criteria.

The assignModule() method selects a nested table of MODULE_T objects from the MODULES column of the PROJECTS table, then updates MODULE_NAME for a particular row of the nested table. Similarly, the deleteUnownedModules() method selects a nested table of MODULE_T objects, then deletes any unowned modules in the nested table, where MODULE_OWNER is null.

These methods use table alias syntax, as discussed previously. In this case, m is used for the nested table, and p is used for the participants table.

/* assignModule 
     Illustrates accessing the nested table using the TABLE construct 
     and updating the nested table row 
  */
  public static void assignModule(int projId, String moduleName, 
                                  String modOwner) throws Exception 
  {
    System.out.println("Update:Assign '"+moduleName+"' to '"+ modOwner+"'");

    try {
      #sql {UPDATE TABLE(SELECT modules FROM projects WHERE id=:projId) m
            SET m.module_owner=
           (SELECT ref(p) FROM participants p WHERE p.ename= :modOwner) 
            WHERE m.module_name = :moduleName };
    } catch(Exception e) {
      System.out.println("Error:insertModules");
      e.printStackTrace();
    }
  }

  /* deleteUnownedModules 
  // Demonstrates deletion of the Nested table element 
  */

  public static void deleteUnownedModules(int projId)
  throws Exception 
  {
    System.out.println("Deleting Unowned Modules for Project " + projId);
    try {
      #sql { DELETE TABLE(SELECT modules FROM projects WHERE id=:projId) m
             WHERE m.module_owner IS NULL };
    } catch(Exception e) {
      System.out.println("Error:deleteUnownedModules");
      e.printStackTrace();
    }
  }

6.6.5 Selecting Data from a Nested Table Using a Nested Iterator

SQLJ supports the use of nested iterators as a way of accessing nested tables. This requires CURSOR syntax, as used in the following example. The code defines a named iterator class, ModuleIter, then uses that class as the type for a modules column in another named iterator class, ProjIter. Inside a populated ProjIter instance, each modules item is a nested table rendered as a nested iterator.

The CURSOR syntax is part of the nested SELECT statement that populates the nested iterators. Once the data has been selected, it is output to the user through the iterator accessor methods.

This example uses required table alias syntax, as discussed previously. In this case, a for the projects table and b for the nested table.

...

//  The Nested Table is accessed using the ModuleIter 
//  The ModuleIter is defined as Named Iterator 

#sql public static iterator ModuleIter(int moduleId , 
                                       String moduleName , 
                                       String moduleOwner);

// Get the Project Details using the ProjIter defined as 
// Named Iterator. Notice the use of ModuleIter:

#sql public static iterator ProjIter(int id, 
                                     String name, 
                                     String owner, 
                                     Date start_date, 
                                     ModuleIter modules);

...

public static void listAllProjects() throws SQLException
{
  System.out.println("Listing projects...");

   // Instantiate and initialize the iterators 

   ProjIter projs = null;
   ModuleIter  mods = null;
   #sql projs = {SELECT a.id, 
                        a.name, 
                        initcap(a.owner.ename) as "owner", 
                        a.start_date,
                        CURSOR (
                        SELECT b.module_id AS "moduleId",
                               b.module_name AS "moduleName",
                                 initcap(b.module_owner.ename) AS "moduleOwner"
                        FROM TABLE(a.modules) b) AS "modules"  
                 FROM projects a };
  
  // Display Project Details
  
  while (projs.next()) {
    System.out.println( "\n'" + projs.name() + "' Project Id:" 
                + projs.id() + " is owned by " +"'"+ projs.owner() +"'"
                + " start on "  
                + projs.start_date());
              
    // Notice the modules from the ProjIter are assigned to the module
    // iterator variable 

    mods = projs.modules();
    System.out.println ("Modules in this Project are : ");

    // Display Module details 

    while(mods.next()) { 
      System.out.println ("  "+ mods.moduleId() + " '"+ 
                                mods.moduleName() + "' owner is '" +
                                mods.moduleOwner()+"'" );
    }                    // end of modules 
    mods.close();
  }                      // end of projects 
  projs.close();
}

6.6.6 Selecting a VARRAY into a Host Expression

This section provides an example of selecting a VARRAY into a host expression. Presume the following SQL definitions:

CREATE TYPE PHONE_ARRAY IS VARRAY (10) OF varchar2(30)
/
/*** Create ADDRESS UDT ***/
CREATE TYPE ADDRESS AS OBJECT
( 
  street        VARCHAR(60),
  city          VARCHAR(30),
  state         CHAR(2),
  zip_code      CHAR(5)
)
/
/*** Create PERSON UDT containing an embedded ADDRESS UDT ***/
CREATE TYPE PERSON AS OBJECT
( 
  name    VARCHAR(30),
  ssn     NUMBER,
  addr    ADDRESS
)
/

CREATE TABLE  employees
( empnumber            INTEGER PRIMARY KEY,
  person_data     REF  person,
  manager         REF  person,
  office_addr          address,
  salary               NUMBER,
  phone_nums           phone_array
)
/

And presume that JPublisher is used to create a PhoneArray custom collection class to map from the PHONE_ARRAY SQL type.

The following method selects a row from this table, placing the data into a host variable of the PhoneArray type:

private static void selectVarray() throws SQLException
{
  PhoneArray ph;
  #sql {select phone_nums into :ph from employees where empnumber=2001};
  System.out.println(
    "there are "+ph.length()+" phone numbers in the PhoneArray.  They are:");

  String [] pharr = ph.getArray();
  for (int i=0;i<pharr.length;++i) 
    System.out.println(pharr[i]);
}

6.6.7 Inserting a Row that Includes a VARRAY

This section provides an example of inserting data from a host expression into a VARRAY, using the same SQL definitions and custom collection class (PhoneArray) as in the previous section.

The following methods populate a PhoneArray instance and use it as a host variable, inserting its data into a VARRAY in the database:

// creates a varray object of PhoneArray and inserts it into a new row
private static void insertVarray() throws SQLException
{
  PhoneArray phForInsert = consUpPhoneArray();
  // clean up from previous demo runs
  #sql {delete from employees where empnumber=2001};
  // insert the PhoneArray object
  #sql {insert into employees (empnumber, phone_nums)
        values(2001, :phForInsert)};
}

private static PhoneArray consUpPhoneArray()
{
  String [] strarr = new String[3];
  strarr[0] = "(510) 555.1111";
  strarr[1] = "(617) 555.2222";
  strarr[2] = "(650) 555.3333";
  return new PhoneArray(strarr);
}

6.7.1 Serializing Java Classes to RAW and BLOB Columns

If you want to store instances of Java classes directly in RAW or BLOB columns, then you must meet certain non-standard requirements to specify the desired SQL-Java mapping. Note that in SQLJ statements the serializable Java objects can be transparently read and written as if they were built-in types.

You have two options in specifying the SQL-Java type mapping:

• Declare a type map in the connection context declaration and use this type map to specify mappings.

• Use the public static final field _SQL_TYPECODE to specify the mapping.

Defining a Type Map for Serializable Classes

Consider an example where SAddress, pack.SPerson, and pack.Manager.InnerSPM, where InnerSPM is an inner class of Manager, are serializable Java classes. In other words, these classes implement the java.io.Serializable interface.

You must use the classes only in statements that use explicit connection context instances of a declared connection context type, such as SerContext in the following example:

SAddress               a =...;
pack.SPerson           p =...;
pack.Manager.InnerSPM pm =...;
SerContext ctx = new SerContext(url,user,pwd,false);
#sql [ctx] { ... :a ... :OUT p ... :INOUT pm ... };

The following is required:

• The connection context type must have been declared using the typeMap attribute of a with clause to specify an associated class implementing java.util.PropertyResourceBundle. In the example, SerContext may be declared as follows.

  #sql public static context SerContext with (typeMap="SerMap");
  
  

• The type map resource must provide non-standard mappings from RAW or BLOB columns to the serializable Java classes. This mapping is specified with entries of the following form, depending on whether the Java class is mapped to a RAW or a BLOB column:

  oracle-class.java_class_name=JAVA_OBJECT RAW
  oracle-class.java_class_name=JAVA_OBJECT BLOB
  
  

  The keyword oracle-class marks this as an Oracle-specific extension. In the example, the SerMap.properties resource file may contain the following entries:

  oracle-class.SAddress=JAVA_OBJECT RAW
  oracle-class.pack.SPerson=JAVA_OBJECT BLOB
  oracle-class.packManager$InnerSPM=JAVA_OBJECT RAW
  
  

  Although the period (.) separates package and class names, you must use the dollar sign ($) to separate an inner class name.

Note that this Oracle-specific extension can be placed in the same type map resource as standard SQLData type map entries.

Using Fields to Determine Mapping for Serializable Classes

As an alternative to using a type map for a serializable class, you can use static fields in the serializable class to determine type mapping. You can add either of the following fields to a class that implements the java.io.Serializable interface, such as the SAddress and SPerson classes from the preceding example:

public final static int _SQL_TYPECODE = oracle.jdbc.OracleTypes.RAW;

public final static int _SQL_TYPECODE = oracle.jdbc.OracleTypes.BLOB;

Limitations on Serializing Java Objects

You should be aware of the effect of serialization. If two objects, A and B, share the same object, C, then upon serialization and subsequent deserialization of A and B, each will point to its own clone of the object C. Sharing is broken.

In addition, note that for a given Java class, you can declare only one kind of serialization: either into RAW or into BLOB. The SQLJ translator can check only that the actual usage conforms to either RAW or BLOB.

RAW columns are limited in size. You might experience run-time errors if the actual size of the serialized Java object exceeds the size of the column.

Column size is much less restrictive for BLOB columns. Writing a serialized Java object to a BLOB column is supported by the Oracle JDBC Oracle Call Interface (OCI) and Thin drivers. Retrieving a serialized object from a BLOB column is supported by all Oracle JDBC drivers since Oracle9i.

Finally, treating serialized Java objects this way is an Oracle-specific extension and requires the Oracle SQLJ run time as well as either the default Oracle-specific code generation (-codegen=oracle during translation) or, for ISO standard code generation (-codegen=iso), Oracle-specific profile customization.

6.7.2 SerializableDatum: an ORAData Implementation

"Additional Uses for ORAData Implementations" includes examples of situations where you might want to define a custom Java class that maps to some oracle.sql.* type other than oracle.sql.STRUCT, oracle.sql.REF, or oracle.sql.ARRAY.

An example of such a situation is if you want to serialize and deserialize Java objects into and out of RAW fields, with a custom Java class that maps to the oracle.sql.RAW type. This could apply equally to BLOB fields, with a custom Java class that maps to the oracle.sql.BLOB type.

This section presents an example of such an application, creating a class, SerializableDatum, that implements the ORAData interface and follows the general form of custom Java classes. The example starts with a step-by-step approach to the development of SerializableDatum, followed by the complete sample code.

1. Begin with a skeleton of the class.

   public class SerializableDatum implements ORAData
   {
      // Client methods for constructing and accessing the Java object
   
      public Datum toDatum(java.sql.Connection c) throws SQLException
      {
         // Implementation of toDatum()
      }
   
      public static ORADataFactory getORADataFactory()
      {
         return FACTORY;
      }
   
      private static final ORADataFactory FACTORY =
              // Implementation of an ORADataFactory for SerializableDatum
   
      // Construction of SerializableDatum from oracle.sql.RAW
   
      public static final int _SQL_TYPECODE = OracleTypes.RAW;
   }
   
   

   SerializableDatum does not implement the ORADataFactory interface, but its getORADataFactory() method returns a static member that implements this interface.

   The _SQL_TYPECODE is set to OracleTypes.RAW because this is the data type being read from and written to the database. The SQLJ translator needs this type code information in performing online type-checking to verify compatibility between the user-defined Java type and the SQL type.

2. Define client methods that perform the following:

   • Create a SerializableDatum object.

   • Populate a SerializableDatum object.

   • Retrieve data from a SerializableDatum object.

   // Client methods for constructing and accessing a SerializableDatum
   
   private Object m_data;
   public SerializableDatum()
   {
      m_data = null;
   }
   public void setData(Object data)
   {
      m_data = data;
   }
   public Object getData()
   {
      return m_data;
   }
   
   

3. Implement a toDatum() method that serializes data from a SerializableDatum object to an oracle.sql.RAW object. The implementation of toDatum() must return a serialized representation of the object in the m_data field as an oracle.sql.RAW instance.

   // Implementation of toDatum()
   
   try {
      ByteArrayOutputStream os = new ByteArrayOutputStream();
      ObjectOutputStream oos = new ObjectOutputStream(os);
      oos.writeObject(m_data);
      oos.close();
      return new RAW(os.toByteArray());
   } catch (Exception e) {
     throw new SQLException("SerializableDatum.toDatum: "+e.toString()); }
   
   

4. Implement data conversion from an oracle.sql.RAW object to a SerializableDatum object. This step deserializes the data.

   // Constructing SerializableDatum from oracle.sql.RAW
   
   private SerializableDatum(RAW raw) throws SQLException
   {
      try {
         InputStream rawStream = new ByteArrayInputStream(raw.getBytes());
         ObjectInputStream is = new ObjectInputStream(rawStream);
         m_data = is.readObject();
         is.close();
      } catch (Exception e) {
        throw new SQLException("SerializableDatum.create: "+e.toString()); }
   }
   
   

5. Implement an ORADataFactory. In this case, it is implemented as an anonymous class.

   // Implementation of an ORADataFactory for SerializableDatum
   
   new ORADataFactory()
   {
      public ORAData create(Datum d, int sqlCode) throws SQLException
      {
         if (sqlCode != _SQL_TYPECODE)
         {
            throw new SQLException
                      ("SerializableDatum: invalid SQL type "+sqlCode);
         }
         return (d==null) ? null : new SerializableDatum((RAW)d);
      }
   };
   
   

6.7.3 SerializableDatum in SQLJ Applications

Given the SerializableDatum class created in the preceding section, this section shows how to use an instance of it in a SQLJ application, both as a host variable and as an iterator column.

Presume the following table definition:

CREATE TABLE PERSONDATA (NAME VARCHAR2(20) NOT NULL, INFO RAW(2000));

SerializableDatum as Host Variable

The following uses a SerializableDatum instance as a host variable:

...
SerializableDatum pinfo = new SerializableDatum();
pinfo.setData (
   new Object[] {"Some objects", new Integer(51), new Double(1234.27) } );
String pname = "MILLER";
#sql { INSERT INTO persondata VALUES(:pname, :pinfo) };
...

SerializableDatum in Iterator Column

Following is an example of using SerializableDatum as a named iterator column:

#sql iterator PersonIter (SerializableDatum info, String name);

...
PersonIter pcur;
#sql pcur = { SELECT * FROM persondata WHERE info IS NOT NULL };
while (pcur.next())
{
   System.out.println("Name:" + pcur.name() + " Info:" + pcur.info());
}
pcur.close();
...

6.7.4 SerializableDatum (Complete Class)

The following is complete code for the SerializableDatum class, which was developed in step-by-step fashion in the preceding sections.

import java.io.*;
import java.sql.*;
import oracle.sql.*;
import oracle.jdbc.*;

public class SerializableDatum implements ORAData
{
// Client methods for constructing and accessing a SerializableDatum

   private Object m_data;
   public SerializableDatum()
   {
      m_data = null;
   }
   public void setData(Object data)
   {
      m_data = data;
   }
   public Object getData()
   {
      return m_data;
   }

// Implementation of toDatum()

   public Datum toDatum(Connection c) throws SQLException
   {

      try {
         ByteArrayOutputStream os = new ByteArrayOutputStream();
         ObjectOutputStream oos = new ObjectOutputStream(os);
         oos.writeObject(m_data);
         oos.close();
         return new RAW(os.toByteArray());
      } catch (Exception e) {
        throw new SQLException("SerializableDatum.toDatum: "+e.toString()); }
   }

   public static ORADataFactory getORADataFactory()
   {
      return FACTORY;
   }

// Implementation of an ORADataFactory for SerializableDatum

   private static final ORADataFactory FACTORY =
   
      new ORADataFactory()
      {
         public ORAData create(Datum d, int sqlCode) throws SQLException
         {
            if (sqlCode != _SQL_TYPECODE)
            {
               throw new SQLException(
                  "SerializableDatum: invalid SQL type "+sqlCode);
            }
            return (d==null) ? null : new SerializableDatum((RAW)d);
         }
      };

// Constructing SerializableDatum from oracle.sql.RAW

   private SerializableDatum(RAW raw) throws SQLException
   {
      try {
         InputStream rawStream = new ByteArrayInputStream(raw.getBytes());
         ObjectInputStream is = new ObjectInputStream(rawStream);
         m_data = is.readObject();
         is.close();
      } catch (Exception e) {
        throw new SQLException("SerializableDatum.create: "+e.toString()); }
   }

   public static final int _SQL_TYPECODE = OracleTypes.RAW;
}

6.8.1 Support for Weakly Typed Objects, References, and Collections

In using Oracle objects, references, or collections in a SQLJ application, you have the option of using generic and weakly typed java.sql or oracle.sql instances instead of the strongly typed custom object, reference, and collection classes that implement the ORAData interface or the strongly typed custom object classes that implement the SQLData interface. Note that if you use SQLData implementations for your custom object classes, then you will have no choice but to use weakly typed custom reference instances.

The following weak types can be used for iterator columns or host expressions in the Oracle SQLJ implementation:

• java.sql.Struct or oracle.sql.STRUCT for objects

• java.sql.Ref or oracle.sql.REF for object references

• java.sql.Array or oracle.sql.ARRAY for collections

In host expressions, they are supported as follows:

• As input host expressions

• As output host expressions in an INTO-list

Using these weak types is not generally recommended, however, as you would lose all the advantages of the strongly typed paradigm that SQLJ offers.

Each attribute in a STRUCT object or each element in an ARRAY object is stored in an oracle.sql.Datum object, with the underlying data being in the form of the appropriate oracle.sql.* subtype of Datum, such as oracle.sql.NUMBER or oracle.sql.CHAR. Attributes in a STRUCT object are nameless. Because of the generic nature of the STRUCT and ARRAY classes, SQLJ cannot perform type checking where objects or collections are written to or read from instances of these classes.

It is generally recommended that you use custom Java classes for objects, references, and collections, preferably classes generated by JPublisher.

6.8.2 Restrictions on Weakly Typed Objects, References, and Collections

A weakly typed object (Struct or STRUCT instance), reference (Ref or REF instance), or collection (Array or ARRAY instance) cannot be used in host expressions in the following circumstances:

• IN parameter if null

• OUT or INOUT parameter in a stored procedure or function call

• OUT parameter in a stored function result expression

They cannot be used in these ways, because there is no way to know the underlying SQL type name, such as Person, which is required by the Oracle JDBC driver to materialize an instance of a user-defined type in Java.