IBM Mainframe Forum

Hi,

I have a small requirement, where i need to replace some values in a sequential file, if a numeric or alphanumeric or alphabetic value is identified.
e.g. if the file have numbers from 01 to 11 for a single record, the numbers will be replaced with 'AA' or if the file have numbers from 12 to 21, the numbers will be replaced by
'BB', or if the file have alphabets ranging from 'A' to 'G', the alphabets will be replaced with '001'

I know it can be achieved through syncsort using ifthen:



but for every single value, I have to write a separate statement like this. Is there any other way to achieve this in more simpler way.. for this small requirement i don't want to use COBOL, and REXX is not allowed here.. so looking for solution in JCL only.

Please give a representative example of input data and expected output.

Thanks Billy..

this is the input value :



and the output will be :



where the replacement is being done as follows:
for numeric: for value from 1 to 11, values replaced with AA
for alphabet: for value from A to G, values replaced with 001, for value from H to L, values replaced with 002, for value from M to T, values replaced with 003

a bit more complicated than it looks

You say values from 1 to 11

if true it means that it must be a two step <parsing>
sometimes substituing a single <char> 1-9
sometime substituting 2 10-11

and after the mangling



and they will be the same ...
is that what You want

it would be wiser to describe better the requirement


and ...

the data does not match the 1 to 11 description

Thanks Enrico,

In case of double digit number in input file.. the format will be as follows:



and the output will be:



In other word, the double digit number will identified using single space after and before( that's y I am also looking for alphanumeric )

I'm not sure this is a "small requirement".

You are translating some "source" value to a "new value" where the "new value" can match a later "source value".

As an example, open a random document in your word processor. Change all "a" to "b". Change all "b" to "c". Is that the result you want? As enrico pointed out, when you look at the "c"s now, you'll have no idea what it was originally.

You are also changing lengths.

You also have a "clash" of source values. "1" exists as "1" and also part of "21", for example.

Unless the intention is to scramble the data beyond use, you have quite a "complex requirement" and it is not even fully defined yet.

Thanks Billy,

Just leave behind the requeirment..
If you can suggest me how to check for numbers ranging from 1 to 11 and replace it with 'AA'.. will serve my purpose by 75%.. rest of the part is logical and functional and I will manage it according to design and requirment.

I know COBOL have this facility with 'THROUGH' verb.. is there anything similar to like this...or some better approach..

Have a look at FINDREP. Specify 11 first, else you'll get AAAA for 11 if you specify 1 first.