FILE-AID to compare a ZS field to a PS field

sandy257 · New User Joined: 03 Feb 2010 Posts: 19 Location: Pune

Hi,

I am trying to compare a ZS field to a PS field on 2 files using batch file-aid compare. But the compare results are showing differences even if the values of both are the same?

Is this a limitation of file-aid compare?
Do I need to convert these fields to a single format before these can be compared?
Is there a shortcut so that file-aid can convert the PS to ZS or vice versa when there is a difference in the format of data? (I am aware of SORT to convert PD to ZD but it will be tedious as no of fields are large).

Background:
I am working on IDMS-DB2 comversion project. Most of the ZS fields in IDMS (some are PS as well) have been converted to decimal format on DB2 (which is PS). Please note I have no control on defining the fields. I have unloaded DB2 and IDMS data onto 2 flat files and comparing them using formatted file-aid compare. Here File-aid is reporting differences on fields whose format are ZS in the IDMS file and PS in the DB2 file even if the values are same.

Nic Clouston · Posted: Wed Jun 15, 2011 6:56 pm

I presume for ZS and PS you mean ZD and PD respectively (respectively Zoned Decimal and Packed Decimal). I believe that compares are usually done byte to byte so, of course, a ZD field would not be the same as a PD field - different lengths for a start. You should run one of the files through a utility (DF/SYNCSORT) to convert the field(s) to match the field(s) of the other file.

Bill Woodger · Posted: Wed Jun 15, 2011 7:06 pm

Maybe the signs are different. If one lot has F and the other C/D?

sandy257 · New User Joined: 03 Feb 2010 Posts: 19 Location: Pune

@Nic.. Yes by ZS and PS I meant Signed ZD and signed PD

I am aware that lengths would differ, but does file-aid not recognize that? One field is S9(10) and other is S9(10) comp-3. So can file-aid not compare these?

I have over 100 pairs of flat files to be compared with several such instances in each of the files. using SORT to do a field wise conversion will be time taking and tedious.

What I was looking at is if file-aid could convert the format before comparing.

Nic Clouston · Posted: Wed Jun 15, 2011 8:22 pm

I do not think I hav ever used File-Aid to compare files - always used SUPERC or COMPAREX both of which do a byte-by-byte comparison not field-by-field comparison. I do not know if File-Aid does a field-by-field compare and eve if it did I do not know if it would normalise ZD v PD fields. First place to look is the manual/F1.

enrico-sorichetti · Posted: Wed Jun 15, 2011 8:29 pm

I am not a fileaid expert, but I guess that with the proper parametrization it can do it

apart the religion wars IBM filemanager and fileaid provide to a large extent the same facilities

and IBM filemanager can do it

Robert Sample · Posted: Wed Jun 15, 2011 8:58 pm

Are you using a LAYOUT for each file?

Studying the example in chapter 7_8 of the File Aid Batch Reference Manual will yield much insight into what is needed to accomplish what you want to do.

sandy257 · New User Joined: 03 Feb 2010 Posts: 19 Location: Pune

Yes I am trying to do a compare using layout for each file. As stated earlier I have 2 files; 1 with IDMS data with IDMS copybook layout and another with DB2 data with DB2 copybook/dclgen layout. I am doing a field by field formatted compare using option 10 of File-aid. When selecting the fields for comparison file-aid is not restricting if I choose to compare a PS field with a ZS field. Generally file-aid provides error messages if field formats of the compared fields are not consistent. I assumed that the content of the field will be compared as PS and ZS fields actually are not much different except for storage lengths. But surprisingly file-aid reported differences on the field even if teh content was the same.

Robert Sample · Posted: Wed Jun 15, 2011 11:08 pm

Since you continue to get basic terminology wrong (ZS and PS are not mainframe data types), and you are not posting any of the control records you ARE using with File Aid, I fail to see where we could possibly help you. Even if we could find a solution for you, your intransigence regarding proper terminology does not bode well for your being able to actually use any soltuion we could provide.

I STRONGLY recommend you open a problem with Compuware and get their assistance since the product is theirs.

Akatsukami · Posted: Wed Jun 15, 2011 11:09 pm

Robert Sample · Posted: Wed Jun 15, 2011 11:19 pm

Akatsukami: perhaps we should just say "ignorance is bliss" in which case sandy257 has to be one of the most blissful people I've ever heard of!

sandy257 · New User Joined: 03 Feb 2010 Posts: 19 Location: Pune

I understand I have been using certain terminologies that are not technically correct but are indicative enough for people of experience to understand.

Anyway I give up. I understand no one has enough experience on this to guide me thru.

enrico-sorichetti · Posted: Wed Jun 15, 2011 11:51 pm

dick scherrer · Posted: Wed Jun 15, 2011 11:57 pm

Hello,

Robert Sample · Posted: Thu Jun 16, 2011 1:21 am

Bill Woodger · Posted: Thu Jun 16, 2011 1:42 am

If you're not going to do anything for yourself on this task, or show us at what point you are stuck, with some information that we might be able to glean something from, at least go to a forum specialising in the generation of apt insults. To accuse the people aiding you of not having "experience" is the most childish come-back I've heard since I was six-years-old.

As for this, come on.

sandy257 · New User Joined: 03 Feb 2010 Posts: 19 Location: Pune

Yes I know they are quite different. The way they store, handle sign storage. But I was under the assumption that file-aid converts/reformats before comparison. I am sorry I assumed a lot of things while posting on this forum.
At the backend file-aid does no more than a character comparison which is why probably the same content is reported as different as had been suggested by someone.

I had already gone thru the file-aid manual before I posted on this forum.

I handled this by using REXX to convert the format to signed ZD. Thanks for all suggestions and brickbats

Bill Woodger · Posted: Thu Jun 16, 2011 11:46 am

When you are in a hole, stop digging. Usually what happens. You brought along extra spades in case you break a few while keeping at it.

Nowhere did you previously indicate that you read the File-Aid manual. You didn't state it, and there is certainly nothing implied in what you wrote.

Now you have used a language which doesn't exactly bother a lot with data-types. If you stick a fish into rexx it will try to operate as though it has a value. Not a good choice to help compare whether good data is coming out of supposedly parallel systems.

Your comparison is now worthless. Toss it in the bin.

Paste some examples of the "matching" data in "code tags" (grey button above input box). Describe how you are doing the comparison.

Or, keep digging.