Overview

The Pervasive PSQL relational interface includes support for COBOL OCCURS constructs, partial REDEFINES, and variable record layouts.

A partial REDEFINE redefines a portion of the data within a record (such as a 05 level within a 01 level). A variable record layout is also referred to as a REDEFINES because the entire record is being redefined. To avoid confusion with terminology, this section refers to partial REDEFINES and to variable record layouts.

No changes are required to your COBOL application to take advantage of the SQL access.

You enable SQL access by describing the application's understanding of data to the Pervasive PSQL relational interface. In developer's terms, you define the metadata to the relational interface.

Note that this section applies only to COBOL applications that include OCCURS, partial REDEFINES, or variable record layouts.

Restrictions

The following restrictions currently apply to providing SQL access for COBOL applications.

OCCURS cannot be nested within OCCURS

OCCURS cannot be nested within partial REDEFINES

Partial REDEFINES cannot be nested within partial REDEFINES

Partial REDEFINES cannot be nested within OCCURS

Only one index can be defined for an OCCURS. Additional indexes cannot be defined for the items within the OCCURS.

The only data types supported are those defined for the Pervasive PSQL transactional and relational interfaces. The data types are described in the XML control file by using only the transactional data types. See cobolschemaexec.xsdin Table D-2 for a discussion of the XML control file. For a discussion of data types, see Pervasive PSQL Supported Data Types in SQL Engine Reference.

SQL Statements

The following table lists the use of SQL statements with data tables created from OCCURS, partial REDEFINES, or variable record layouts.

Table D-1 Use of SQL Statements With COBOL Data Constructs
Statement	Use with OCCURs and partial REDEFINES	Use with Variable Record Layouts	Notes
ALTER TABLE	No	No
CREATE INDEX	No	No
CREATE INDEX IN DICTIONARY	No	Yes
CREATE TRIGGER	No	No
DELETE	No	Yes
DROP TABLE	Yes	Yes	A DROP TABLE statement removes all of the entries from the system tables. The data file itself is not deleted or modified. Also, when you drop a main table, a message informs you to drop any dependent tables if any are detected. A dependent table depends on a main table and results from conditions such as an OCCURS that contains an index or from partial REDEFINES. Once you drop the dependent tables, you can drop the main table.
INSERT INTO	No	No
UPDATE	Yes	Yes	An UPDATE statement cannot update a column on which a table filter has been defined. A table filter is a logical expression associated with a table. Table filters are defined as part of your metadata in the XML files.
all other SQL statements listed in SQL Engine Reference.	Yes	Yes	See Grammar Statements in SQL Engine Reference.