Comments on: How to handle Null value in COBOL/400.

By: The Most-Watched IT Questions: May 24, 2011 - ITKE Community Blog

The Most-Watched IT Questions: May 24, 2011 - ITKE Community Blog — Tue, 24 May 2011 06:41:36 +0000

[...] A member wondering how to handle null value in COBOL/400 is getting some help from TomLiotta, CharlieBrowne, and MDratwa. Jump in with your own expertise! [...]

By: tomliotta

tomliotta — Thu, 12 May 2011 22:40:17 +0000

Flat files (unkeyed) can not contain nulls (null capable fields). You must use keyed PF or keyed LF:

Flat files do not equate with ‘unkeyed’. Unkeyed files are not necessarily ‘flat files’ even though flat files will be unkeyed.

Further, null-capable columns can exist in unkeyed files. I have a number of them.

However, thanks for pointing out that this might not be a SQL issue. It hasn’t actually been stated if the COBOL is using SQL or native I/O to process the file. The way that the term “view” was used in the question put me into thinking that SQL was the focus.

Your native I/O example could be the right direction.

Tom

By: mdratwa

mdratwa — Thu, 12 May 2011 20:52:00 +0000

Flat files (unkeyed) can not contain nulls (null capable fields). You must use keyed PF or keyed LF: Null key map is used when any of the key fields are null capable. Null map is used when any of the fields (key or not) are null capable. In the null map or null key map fields will contain a B'1' or B'0' ('0' = not null, '1' = null). Before any updating (write/rewrite), if you change a null to a value other than null, the corr null map field (and/or null key map) for that field needs to be changed from B'1' to B'0'. The source must be changed from "CBL" to "CBLLE". SELECT IOF-GIS ASSIGN DATABASE-XXXLFA40W-ALWNULL ORGANIZATION INDEXED ACCESS DYNAMIC RECORD KEY EXTERNALLY-DESCRIBED-KEY WITH DUPLICATES FILE STATUS WS-FILE-STATUS. Working Storage Section. 01 WS-NULL-FILE-AREA. COPY DDS-ALL-FORMATS OF XXXLFA40W WITH NULL-MAP WITH NULL-KEY-MAP. Procedure Division. READ IOF-GIS NULL-KEY-MAP IS SOMPFA40W-NKM NULL-MAP IS SOMPFA40W-NM INVALID KEY (or next record) go to xxxx END-READ.

By: tomliotta

tomliotta — Wed, 11 May 2011 22:02:37 +0000

If your tables have null-capable columns and any of those columns have null values, your program must use null-indicators in order to fetch those rows successfully.

Tom

By: rr22

rr22 — Wed, 11 May 2011 11:17:10 +0000

Thanks for the reply.It is in COBOL/400.I am not using the null values in our program.It is present in View.When i am reading the view,it is reading same record due to Null value & giving 46 error code.

By: tomliotta

tomliotta — Mon, 09 May 2011 21:11:16 +0000

The SQL Programming with Host Languages manual has examples for COBOL coding. As with RPG, null columns will supply a “null indicator” that you need to test before using a variable that has a null value. An example might be like the following.

Given the statement:

EXEC SQL FETCH CLS_CURSOR INTO :CLS-CD,
                               :NUMDAY :NUMDAY-IND,
                               :BGN :BGN-IND,
                               :ENDCLS :ENDCLS-IND
END-EXEC.

The variables can be declared as follows:

EXEC SQL BEGIN DECLARE SECTION END-EXEC.
77 CLS-CD     PIC X(7).
77 NUMDAY     PIC S9(4) BINARY.
77 BGN        PIC X(8).
77 ENDCLS     PIC X(8).
77 NUMDAY-IND PIC S9(4) BINARY.
77 BGN-IND    PIC S9(4) BINARY.
77 ENDCLS-IND PIC S9(4) BINARY.
EXEC SQL END DECLARE SECTION END-EXEC.

Is that what you’re looking for?

Tom

By: tomliotta

tomliotta — Mon, 09 May 2011 20:05:29 +0000

Also, can you clarify if this is COBOL/400 (OPM) or ILE COBOL? The program code might be the same, but the compiled behaviors may be slightly different.

Tom

By: charliebrowne

charliebrowne — Mon, 09 May 2011 18:32:55 +0000

Can you post the code.
If you are just doing a read, it should not make any difference if a value is null.
If you are reading an LF and updating, then there is more of a possibility to keep on reading the same record.