Alpha Numeric

[Angu85]

Hi

I am working in online programing. As we send the response some time

we send a junk data Invalid alphanumeric . We have log facilities but still

it is hard to find unless we do an attended testing . Is there any way where

we can check the response before we send . Just like

If Numeric . Note that the PIC of response we send is X(200000) . Please

Suggest an efficient way to check this out .

[Anuj Dhawan]

I'm lost - unfortunately the description of your problem is not clear. Can you please start over again.

[Natarajan]

Based my understanding, your question is how we can stop sending junk data back to front end / calling program.

Before populating output data fields.. initialize those items. This will take care of problem of junck being sent back to calling program/front end.

If my understanding your problem is not correct. give clear details about your problem.

[dbzTHEdinosauer]

CEDF would be a solution
explanation of CEDF
using CEDF

you have provided us with no information except to use non-technical terms (junk data)
to describe a relatively beginner-type problem - not insuring the integrity of your outbound data.
Natarajan provided a start.
The size of your outbound-data area, X(200000), indicates that you are not sending to a screen.

If you have at your site 3rd party debugging tools, you could also use them.

if you reply provide a little more detail, as Anuj and Natarajan have requested.
otherwise their expertise is of no use.

I am also moving this question to the CICS forum, where it belongs.

[Angu85]

Ok Let me explain the question again. We fetch data for back end file and send it to front end . Out

of thousands of fields which we extract from the file 1 seems to be unprintable character in the file.

i.e. a X(20) which looks like this X'0b12fds3j3uj45k6j7kl' ( hexadecimal value ) . Now this X(20)

is sent along with X(20000) . i.e. X(20) is one of the fields among X(20000) .

* Here is scnerio we sent crores of responses like X(20000) every day . Out of which I want to

Identify the responses of above type ( with unprintable characters ) and write it in a separate log.

file.

Is there any efficient way to find such response . This has to be efficient since we cant delay

in sending the response in online.

[Angu85]

Ok Let me explain the question again. We fetch data for back end file and

send it to front end . Out of thousands of fields which we extract from the

file, 1 seems to be unprintable character in the file.

i.e. a X(20) which looks like this X'0b12fds3j3uj45k6j7kl' ( hexadecimal

value ).Now this X(20) is sent along with X(20000) . i.e. X(20) is one of

the fields among X(20000) .

* Here is scnerio we sent crores of responses like X(20000) every day .

Out of which I want to Identify the responses of above type ( with

unprintable characters ) and write it in a separate log file.

Is there any efficient way to find such response . This has to be efficient

since we cant delay in sending the response in online.

[dbzTHEdinosauer]

Angu85.

X'0b12fds3j3uj45k6j7kl' is not hexadecimal. 0-9, A-F, nothing else.

you need to decide what are 'invalid' characters.

you can use INSPECT Tallying to determine if there are any invalid char
and INSPECT CONVERT to change the data in a field containing invalid char.

if all you want to do is replace these 'invalid' characters with spaces,
INSPECT convert (format 4) , example

you can set-up your translate tables with hex values.

you could also have 1 validation routine that you would perform on a common field,
which you would populate with the field in question.
that would save coding many INSPECT statements.

you could redefine the common-field as a table,
use level 88's on the indexed char to determine invalid-char
as well as setting the value of the 'bad' char to space.
a well written perform varying index will out perform an INSPECT.

The best way to solve this problem is to correct the invalid data at source,
not during an online transaction.

[Angu85]

Thanks a lot for your patient answer Guys .