Database Bridge generator (edbiis)

The EDBIIS command allows the user to generate EDBI routines for supported RDBMS.

This command is mainly supported for backward compatibility. If you activated the Database Bridge as described in EDBI Generation at compile time (one step), then you don’t need to use this command.

Instead, if you wish to generate EDBI routines in a separate step by processing EFD dictionaries, then you need this command.

Syntax

edbiis -help|[options] <efdfilename>

Command Line Options

(default)	Generates EDBI routines for generic RDBMS.
-d2	Generates EDBI routines for DB2 RDBMS.
-d4	Generates EDBI routines for DB2/AS400 RDBMS.
-di	Generates EDBI routines for Informix (certified for ANSI-mode databases).
-do	Generates EDBI routines for ORACLE RDBMS.
-dm	Generates EDBI routines for MySQL (InnoDB engine).
-dmld	Generates EDBI routines with light cursors for MySQL (InnoDB engine). See iscobol.easydb.mysql_row_limit for details.
-dmlu	Generates EDBI routines with light cursors for unique indexes for MySQL (InnoDB engine). See iscobol.easydb.mysql_row_limit for details.
-dmoid=<name>	Specifies the name of the OID field generated when -dmld is used. The default name is "OID".
-dp	Generates EDBI routines for PostgreSQL.
-dpld	Generates EDBI routines with light cursors for PostgreSQL. See iscobol.easydb.postgres_row_limit for details.
-dplu	Generates EDBI routines with light cursors for unique indexes for PostgreSQL. See iscobol.easydb.postgres_row_limit for details.
-ds	Generates EDBI routines for Microsoft SQL Server.
For every database:
-ca	Uses A4GL-WHERE-CONSTAINT instead of The EDBI-WHERE-CONSTRAINT external variable and implies -defCHAR.
-cc	Includes the COMMIT COUNT feature in the EDBI routine. A COMMIT statement is automatically performed after a given number of successful WRITE, REWRITE and DELETE operations. The number is configured by the properties iscobol.easydb.commit_count.ConnectionName and iscobol.easydb.commit_count.   Note that the COMMIT COUNT feature requires iscobol.jdbc.autocommit (boolean) to be set to false in the configuration, otherwise all the operations are automatically committed.
-ce	Generate direct SELECT query instead of using a CURSOR to perform READ KEY.
-cp	Generates subroutines that are suitable to be called only by COBOL programs compiled with -cp option. Without this option, the subroutines can be called only by COBOL programs compiled without -cp option. When -cp is used, the following additional steps must be performed in order to work in cp mode: • set iscobol.file.index=easydbcp in the configuration • compile the EDBI subroutines with -cp option as well
-defCHAR	Manages alphanumeric fields using CHAR instead of VARCHAR.   Use the VAR-LENGTH Directive to mark specific alphanumeric fields as VARCHAR.
-entrypoints	Generates entry points where the user can inject customized code.   See Extending EDBI routines through entry points for more information.
-esst	Generates additional code to provide the ability to use only the table related to a specific record type in multi-record files. To enable the feature at runtime, set iscobol.easydb.start_on_specific_table (boolean) to true in the configuration.
-h	Uses HIGH-VALUE as max numeric value in numeric key fields.   This option affects numeric items that cannot be set to HIGH-VALUE. It doesn’t affect COMP, BINARY, COMP-X, COMP-5 and COMP-2.
-i	ISAM positioning on at end   Using the -i option produces a different behavior when reversing direction after reading past the beginning or end of a file.   The record returned by the READ PREVIOUS is the second-to-last record in the file, and the record returned by the READ NEXT is the second record in the file.
-jcd=<routine_name>	Specifies an alternate routine for the conversion of julian dates before writing on the database.   By default EDBI_DTJUCBDB (installed with the product) is used. This routine takes advantage of the DATE-OF-INTEGER intrinsic function for the conversion. If you wish to write your own routine that uses a different conversion logic, use the same Linkage parameters as EDBI_DTJUCBDB.CBL found in easydb\edbisource in the isCOBOL installation.   The custom routine you provide is searched in the iscobol.code_prefix, if set, or in the Class Path otherwise.   The -jcd option must be used in conjunction with -jdc.
-jdc=<routine_name>	Specifies an alternate routine for the conversion of julian dates after reading from the database.   By default EDBI_DTJUDBCB (installed with the product) is used. This routine takes advantage of the INTEGER-OF-DATE intrinsic function for the conversion. If you wish to write your own routine that uses a different conversion logic, use the same Linkage parameters as EDBI_DTJUDBCB.CBL found in easydb\edbisource in the isCOBOL installation.   The custom routine you provide is searched in the iscobol.code_prefix, if set, or in the Class Path otherwise.   The -jdc option must be used in conjunction with -jcd.
-ld=<dirname>	Specifies the directory in which template files are stored. Edbiis needs template files to generate the routines.
-maxCHARlen=n	Alphanumeric fields whose size is not greater than n are managed as CHAR, the others are managed as VARCHAR.   This option is not compatible with -defCHAR and overrides both FIX-LENGTH Directive and VAR-LENGTH Directive.
-mo	Generates multitable subroutines using standard COBOL statements instead of object oriented syntax.
-n	Test not numeric   Using the -n option will include an additional test on numeric fields to verify whether a numeric value is used. This additional check will determine if non-numeric values are used and replace those non-numeric values with 0.
-no	Use leading zeroes in OCCURS item names. The number of leading zeroes depends by the occurs size. EasyDB puts before as many zeroes as it takes to reach the number of digits of the occurs size.   Example: Consider the following COBOL items: 03 my_item_a pic x(10) occurs 3. 03 my_item_b pic x(10) occurs 30. 03 my_item_c pic x(10) occurs 300. Without -no option the columns generated by edbiis are named: my_item_a_1, my_item_a_2, my_item_a_3 my_item_b_1, my_item_b_2, my_item_b_3, ... my_item_b_30 my_item_c_1, my_item_c_2, my_item_c_3, ... my_item_c_300 With -no option the columns are named: my_item_a_1, my_item_a_2, my_item_a_3 my_item_b_01, my_item_b_02, my_item_b_03, ... my_item_b_30 my_item_c_001, my_item_c_002, my_item_c_003, ... my_item_c_300
-oc=<dirname>	Output directory for copyfile.
-od=<dirname>	Output directory for EDBI routines.
-os=<dirname>	Output directory for sql script.
-pdo	List records with duplicate keys values ordered by the primary key during READ NEXT and READ PREVIOUS on alternate keys with duplicates. Without this option there is no guarantee to read the correct record by reading on the opposite way when the file pointer is on a duplicated key value. This option might slow down performance.
-sl	Support for START WITH SIZE. Using this option, edbiis stores additional code in the routine to handle the SIZE clause of the START statement. If the -sl option is omitted, the routine will handle the START statement as if the SIZE clause is not specified.
-t	Allow trace for not numeric   The -t option must be used along with -n. Routines generated with -t and -n options keep trace of cases of not numeric data in numeric field in a separate log file whose name is controlled by the iscobol.edbi.notnum.tracefile configuration property.
-ua	Generate additional code to provide support for the statement UNLOCK file-name ALL RECORDS. The statement UNLOCK ALL is always ignored, instead. Warning: In order to unlock the file, the EDBI routine will issue a COMMIT statement, so every active lock will be lost.
-v	Show product version.
Only for Informix:
-Id	Use strings to represent date values in SQL statements. Avoid conversion functions. This is useful when working with old Informix versions where date conversion functions were not available.
Only for Oracle:
-Oh	Generate Oracle optimizer hints that force the query optimizer to use the proper index. This option is deprecated and supported only for backward compatibility. Unless the HINT Directive has been used to specify custom hints, you may consider using the new option -Oho.
-Oho	Generate Oracle optimizer hints that force the query optimizer to use the proper index. Hints are also used to specify the data ordering, avoiding the Order By clause and providing better performance. The HINT Directive shouldn't be used along with this option since it might cause wrong data ordering. This option is incompatible with the -Oh option and with the configuration setting iscobol.jdbc.cursor.type=3.
-Oii=<integer>	Initial storage value for index.
-Oit=<integer>	Initial storage value for table.
-Oni=<integer>	Next storage value for index.
-Ont=<integer>	Next storage value for table.
-Opi=<integer>	pctincrease storage value for index.
-Opt=<integer>	pctincrease storage value for table.
-Oti=<name>	Tablespace index name.
-Ott=<name>	Tablespace name.
-Ow	NOWAIT for update. This option allows the EDBI user's routine to return record lock condition.
-Owfl	Includes the WAIT FOR LOCKS feature in the EDBI routine. Before each READ operation the EDBI routine tests the iscobol.easydb.wait_for_lock (boolean) configuration property. If the property is set to true, then the lock condition is not returned and the program waits for the lock to be released. If the property is set to false, then the lock condition is returned. This option can’t be used along with -Ow.
Only for Ms SQL Server:
-Sc	Use standard ASCII comparison when sorting data.   By default, without this option, in collation comparisons that use Windows collations, characters like a single quote (') or hyphen (-) are compared last, only after the regular alphabet characters are compared.   With this option COLLATE Latin1_General_Bin is used upon CREATE TABLE.
-Sco	Use standard ASCII comparison when sorting data.   By default, without this option, in collation comparisons that use Windows collations, characters like a single quote (') or hyphen (-) are compared last, only after the regular alphabet characters are compared.   With this option COLLATE Latin1_General_Bin is used in the ORDER BY clause of queries.
-Sdt	Use always DATETIME to represent COBOL fields with the EFD DATE directive, regardless of the date format string.
Only for PostgreSQL:
-Pi	Use indicator variables to manage COBOL Low-Values as NULL on the database.

Unlike other drivers, the PostgreSQL JDBC driver tries to load all the records of a Cursor (object used by EDBI subroutines to store table records ) into memory. For this reason, when the program performs a START on a huge table, an out of memory error may occur. The -dpld and -dplu options help to avoid this situation. When used, the EDBI subroutine will include a pagination logic that keeps the Cursor light. Use one of these options instead of -dp if you plan to work on huge tables with PortgreSQL.