2 Teradata Gateway Features and Restrictions

After the gateway is installed and configured, you can use the gateway to access Teradata data, pass Teradata commands from applications to the Teradata database, perform distributed queries, and copy data.

This chapter contains the following sections:

2.1 Using the Pass-Through Feature

The gateway can pass Teradata commands or statements from the application to the Teradata database using the DBMS_HS_PASSTHROUGH package.

Use the DBMS_HS_PASSTHROUGH package in a PL/SQL block to specify the statement to be passed to the Teradata database, as follows:

DECLARE
    num_rows INTEGER;
BEGIN
    num_rows := DBMS_HS_PASSTHROUGH.EXECUTE_IMMEDIATE@TERA('command');
END;
/

Where command cannot be one of the following:

  • BEGIN TRANSACTION

  • BT

  • COMMIT

  • END TRANSACTION

  • ET

  • ROLLBACK

The DBMS_HS_PASSTHROUGH package supports passing bind values and executing SELECT statements.

See Also:

Oracle Database PL/SQL Packages and Types Reference and Chapter 3, Features of Oracle Database Gateways, of Oracle Database Heterogeneous Connectivity Administrator's Guide for more information about the DBMS_HS_PASSTHROUGH package.

2.2 Database Compatibility Issues for Teradata

Teradata and Oracle databases function differently in some areas, causing compatibility problems. The following compatibility issues are described in this section:

2.2.1 Schema Considerations

The Oracle concept of a schema does not exist in Teradata. An owner included in a query is recognized by Teradata as a database name. In the Oracle database, the schema of an object is identical to the owner of that object. However, when one retrieves or references an OWNER field in a Data Dictionary view such as ALL_TABLES, the OWNER field really is referencing a Teradata database.

When querying data dictionary tables, the following results are returned:

  • ALL_* data dictionary tables, data for every Teradata database is returned. The Teradata database name is returned in the OWNER column.

  • ALL_USERS data dictionary table, the name of every Teradata database is returned.

  • USER_* data dictionary tables, data for the default Teradata database is returned. If a default Teradata database is not defined, the DBC Teradata system database is used.

2.2.2 Naming Rules

Naming rule issues include the following:

2.2.2.1 Rules for Naming Objects

Oracle and Teradata use different database object naming rules. For example, the maximum number of characters allowed for each object name can be different. Also, the use of single and double quotation marks, case sensitivity, and the use of alphanumeric characters can all be different.

See Also:

Oracle Database Reference and Teradata documentation.

2.2.2.2 Case Sensitivity

The Oracle database defaults to uppercase unless you surround identifiers with double quote characters. For example, to refer to the Teradata table called emp, enter the name with double quote characters, as follows:

SQL> SELECT * FROM "emp"@TERA;

However, to refer to the Teradata table called emp owned by Scott from an Oracle application, enter the following:

SQL> SELECT * FROM "Scott"."emp"@TERA;

If the Teradata table called emp is owned by SCOTT, a table owner name in uppercase letters, you can enter the owner name without double quote characters, as follows:

SQL> SELECT * FROM SCOTT."emp"@TERA;

or

SQL> SELECT * FROM scott."emp"@TERA;

Oracle recommends that you surround all Teradata object names with double quote characters and use the exact letter case for the object names as they appear in the Teradata data dictionary. This convention is not required when referring to the supported Oracle data dictionary tables or views listed in Appendix C, "Data Dictionary".

If existing applications cannot be changed according to these conventions, create views in Oracle to associate Teradata names to the correct letter case. For example, to refer to the Teradata table emp from an existing Oracle application by using only uppercase names, define the following view:

SQL> CREATE VIEW EMP (EMPNO, ENAME, SAL, HIREDATE)      AS SELECT "empno", "ename", "sal", "hiredate"      FROM "emp"@TERA;

With this view, the application can issue statements such as the following:

SQL> SELECT EMPNO, ENAME FROM EMP;

Using views is a workaround solution that duplicates data dictionary information originating in the Teradata data dictionary. You must be prepared to update the Oracle view definitions whenever the data definitions for the corresponding tables are changed in the Teradata database.

2.2.3 Data Types

Data type issues include the following:

2.2.3.1 Binary Literal Notation

Oracle SQL uses hexadecimal digits surrounded by single quotes to express literal values being compared or inserted into columns defined as data type RAW.

This notation is not converted to syntax compatible with the Teradata VARBINARY and BINARY data types (a ff surrounded by single quotes followed by hexadecimal digits).

For example, the following statement is not supported:

SQL> INSERT INTO BINARY_TAB@TERA VALUES ('ff'xb)

Where BINARY_TAB contains a column of data type VARBINARY or BINARY. Use bind variables when inserting into or updating VARBINARY and BINARY data types.

2.2.3.2 Data Type Conversion

Teradata does not support implicit date conversions. Such conversions must be explicit.

For example, the gateway issues an error for the following SELECT statement:

SELECT DATE_COL FROM TEST@TERA WHERE DATE_COL = "1-JAN-2001";

To avoid problems with implicit conversions, add explicit conversions, as in the following:

SELECT DATE_COL FROM TEST@TERA WHERE DATE_COL = TO_DATE("1-JAN-2001")

See Also:

Appendix A, "Data Type Conversion" for more information about restrictions on data types.

2.2.4 Queries

Query issues include the following:

2.2.4.1 Row Selection

Teradata evaluates a query condition for all selected rows before returning any of the rows. If there is an error in the evaluation process for one or more rows, no rows are returned even though the remaining rows satisfy the condition.

Oracle evaluates the query condition row-by-row and returns a row when the evaluation is successful. Rows are returned until a row fails the evaluation.

2.2.4.2 Empty Bind Variables

For VARCHAR bind variables, the gateway passes empty bind variables to the Teradata database as a NULL value.

2.2.5 Locking

The locking model for an Teradata database differs significantly from the Oracle model. The gateway depends on the underlying Teradata behavior, so Oracle applications that access Teradata through the gateway can be affected by the following possible scenarios:

  • Read access might block write access.

  • Write access might block read access.

  • Statement-level read consistency is not guaranteed.

    See Also:

    Teradata documentation for information about the Teradata locking model.

2.3 Known Restrictions

If you encounter incompatibility problems not listed in this section or in "Known Problems", contact Oracle Support Services. The following section describes the known restrictions and includes suggestions for dealing with them when possible:

2.3.1 Transactional Integrity

The gateway cannot guarantee transactional integrity in the following cases:

  • When a statement that is processed by the gateway causes an implicit commit in the target database

  • When the target database is configured to work in autocommit mode

    Note:

    Oracle strongly recommends the following:
    • If you know that executing a particular statement causes an implicit commit in the target database, then ensure that this statement is executed in its own transaction.

    • Do not configure the target database to work in autocommit mode.

2.3.2 Transaction Capability

The gateway does not support savepoints. If a distributed update transaction is under way involving the gateway and a user attempts to create a savepoint, the following error occurs:

ORA-02070: database dblink does not support savepoint in this context

By default, the gateway is configured as COMMIT_CONFIRM and it is always the commit point site when the Teradata database is updated by the transaction.

2.3.3 COMMIT or ROLLBACK in PL/SQL Cursor Loops Closes Open Cursors

Any COMMIT or ROLLBACK issued in a PL/SQL cursor loop closes all open cursors, which can result in the following error:

ORA-1002:  fetch out of sequence 

To prevent this error, move the COMMIT or ROLLBACK statement outside the cursor loop.

2.3.4 Pass-Through Feature

Oracle recommends that you place a DDL statement in its own transaction when executing such a statement with the pass-through feature. An explicit COMMIT must be issued after the DDL statement.

If the SQL statements being passed through the gateway result in an implicit commit at the Teradata database, the Oracle transaction manager is unaware of the commit and an Oracle ROLLBACK command cannot be used to roll back the transaction.

2.3.5 Bind Variables for Date Columns

You cannot compare columns of data type TIME or TIMESTAMP to a bind variable.

The following SQL statement causes an error message:

SQL> select time_column from time_table@TERA where time_column = :a;

The following error is issued:

Invalid operation on an ANSI Datetime or Interval value. 

2.3.6 SQL Syntax

This section lists restrictions on the following SQL syntax:

See Also:

Appendix B, "Supported SQL Syntax and Functions" for more information about restrictions on SQL syntax.

2.3.6.1 WHERE CURRENT OF Clause

UPDATE and DELETE statements with the WHERE CURRENT OF clause are not supported by the gateway because they rely on the Oracle ROWID implementation. To update or delete a specific row through the gateway, a condition style WHERE clause must be used.

2.3.6.2 CONNECT BY Clause

The gateway does not support the CONNECT BY clause in a SELECT statement.

2.3.6.3 ROWID

The Oracle ROWID implementation is not supported.

2.3.6.4 EXPLAIN PLAN Statement

The EXPLAIN PLAN statement is not supported.

2.3.6.5 Callback Support

SQL statements that require the gateway to callback to Oracle database would not be supported.

The following categories of SQL statements will result in a callback:

  • Any DML with a sub-select, which refers to a table in Oracle database. For example:

    INSERT INTO emp@TERA SELECT * FROM oracle_emp;
    
  • Any DELETE, INSERT, UPDATE or "SELECT... FOR UPDATE..." SQL statement containing SQL functions or statements that need to be executed at the originating Oracle database.

    These SQL functions include USER, USERENV, and SYSDATE, and the SQL statements are in selects of data from the originating Oracle database. For example:

    DELETE FROM emp@TERA WHERE hiredate > SYSDATE;
    
    SELECT ename FROM tkhoemp@TERA
    WHERE hiredate IN (SELECT hiredate FROM oracle_emp)
    FOR UPDATE OF empno;
    
  • Any SQL statement that involves a table in Oracle database, and a LONG or LOB column in a remote table. For example:

    SELECT a.long1, b.empno FROM scott.table@TERA a, emp b
    WHERE a.id=b.empno; 
    
    SELECT a.long1, b.dummy FROM table_non@TERA a, dual b;
    

where a.long1 is a LONG column.

2.3.7 SQL*Plus

You need to use double quotes to wrap around lowercase table names.

For example:

copy from tkhouser/tkhouser@inst1 insert loc_tkhodept using select * from "tkhodept"@TERA;

2.3.8 Database Links

The gateway is not a shared server process and cannot support shared database links. Each gateway session spawns a separate gateway process and connections cannot be shared.

2.3.9 Stored Procedures

The gateway does not support the procedure feature that allows the execution of stored procedures in a non-Oracle database.

2.4 Known Problems

This section describes known problems and includes suggestions for correcting them when possible. If you have any questions or concerns about the problems, contact Oracle Support Services. A current list of problems is available online. Contact your local Oracle office for information about accessing the list.

The following known problems are described in this section:

2.4.1 Encrypted Format Login

The Oracle database no longer supports the initialization parameter DBLINK_ENCRYPT_LOGIN. Up to version 7.3, this parameter's default TRUE value prevented the password for the login user ID from being sent over the network (in the clear). Later versions automatically encrypt the password.

If this parameter is set to TRUE in the initialization parameter file used by the Oracle9i database, you must change the setting to FALSE, the default setting, to allow Oracle9i to communicate with the gateway.

In the current release, Oracle Database 11g, Release 11.1, the DBLINK_ENCRYPT_LOGIN initialization parameter is obsolete, so you need not check it.

2.4.2 Teradata LONG VARCHAR Data Type

The following restrictions apply when using LONG VARCHAR data types:

  • An unsupported SQL function cannot be used in a SQL statement that accesses a column defined as Teradata data type LONG VARCHAR.

  • You cannot use SQL*Plus to select data from a column defined as Teradata data type LONG VARCHAR when the data is greater than 80 characters in length. Oracle recommends using Pro*C or Oracle Call Interface to access such data in a Teradata database.

  • LONG VARCHAR data types must be NULLABLE for INSERT or UPDATE to work.

  • A table including a LONG VARCHAR column must have a unique index defined on the table or the table must have a separate column that serves as a primary key.

  • LONG VARCHAR data cannot be read through pass-through queries.

The gateway does not support the PL/SQL function COLUMN_VALUE_LONG of the DBMS_SQL package.

See Also:

Appendix B, "Supported SQL Syntax and Functions" for more information about restrictions on SQL syntax.

2.4.3 Schema Names and PL/SQL

If you do not prefix a Teradata database object with its schema name in a SQL statement within a PL/SQL block, the following error message occurs:

ORA-6550 PLS-201 Identifier table_name must be declared.

Change the SQL statement to include the schema name of the object.

2.4.4 Data Dictionary Views and PL/SQL

You cannot refer to data dictionary views in SQL statements that are inside a PL/SQL block.