TRT saving

Pankaj Shrivastava · New User Joined: 24 Jul 2009 Posts: 51 Location: Pune

Hi ,

Is there any way we can avoid declaring a table of 256 byte in TRT when i need to check a byte for Numeric only .

thanks

Bill O'Boyle · Posted: Wed Aug 05, 2009 3:49 pm

You can check the one-byte (using a CLI) for a value not less than X'F0' (C'0') and not greater than X'F9' (C'9').

I'm assuming you're talking about display-numeric data and not packed-decimal data?

Is this CICS/Assembler?

Bill

Pankaj Shrivastava · New User Joined: 24 Jul 2009 Posts: 51 Location: Pune

Hi Bill ,

I have to go for TRT , sorry my question appeared as if i have one byte only ..I have to check each byte in a string for numeric . i.e I have 20 byte of string and need to check if is pure numeric .

Thanks

Bill O'Boyle · Posted: Wed Aug 05, 2009 4:06 pm

Define your table as 256X'04', then ORG table+240 with XL10'00'.

If the field contains all numerics, then a CC "BZ" will be raised after the TRT.

Otherwise, R1 will contain the address of the bad-byte and R2 will contain the actual byte-value found.

Optionally, you can clear R1 and R2 to X'00's before the TRT, but it's not necessary.

Some shops have pre-defined tables, which you can bring into the program, via an MVS LOAD Macro (Batch) or CICS LOAD PROGRAM (CICS). It saves base-register addressability and allows a common table for validation, available to all programs.

Bill

Bill O'Boyle · Posted: Wed Aug 05, 2009 6:38 pm

Optionally, you could use a BCT loop. R14 and R15 are used in this example and should not be a problem. R1 will contain the number of non-numeric bytes found.

Pankaj Shrivastava · New User Joined: 24 Jul 2009 Posts: 51 Location: Pune

Thx Bill , The code is very useful .