MVSFORUMS.com

shuko · Beginner Joined: 08 Nov 2005 Posts: 73 Topics: 20

I need to compare 2 files (both LRECL=80 and FB for this example)
and check the values in file1 with file2. The key is in postion 15
This check has to be done for all columns starting with postion 45
I will have about 45 columns. For convenience sake I have just shown
a few.
My idea was to use SELECT FROM ... ON ... NODUPS. Since SELECT only
permits 10 ON's, this wont work for me.
Is there a better way of doing this?. Any change in any columns
for a key,starting position 45,should cause the record in
file1 with reversed amount written to file3 and file2 record
also written to file3. My problem now is how can I check
several columns and write out the result.

An Example

kolusu · Posted: Tue Feb 21, 2006 8:15 am Post subject:

Shuko,

Couple of questions.

1. Do you have any duplicates in any of the files?

2. What happens when there is no matching record in the file2 or vice-versa?

3. what version of sort do you have (DFSORT , syncsort ....)

Kolusu
_________________
Kolusu
www.linkedin.com/in/kolusu

shuko · Beginner Joined: 08 Nov 2005 Posts: 73 Topics: 20

kolusu,

1. The same files will not have duplicates, if any duplicates found
across files, they can be discarded.
2. If file1 record not found in file2, discard this record and if file2 record
not found in file1, this should appear in out file.
3 I have Z/OS DFSORT V1R5

On second thoughts, is it possible to have two output files one with
file1 records and the second with file2 records based on the following
criteria.

If file1 keys not in file2, discard file1 record
If file2 keys not in file1, write file2 records to OUT2
If file2 keys found in file1 and any column values starting with position
45 are different, then file1 record in OUT1 and file2 record to OUT2
If file2 keys found in file1 and all the col. values are the same starting
with position 45, then discard file1 and file2 records

Frank Yaeger · Posted: Tue Feb 21, 2006 11:37 am Post subject:

Are you asking now how to do your "second thoughts" requirement? If so, please show an example of the records in each input file and the records you expect for output, illustrating each of your conditions.

I don't understand your problem with the limit of 10 ON fields. A CH ON field can have up to 1500 bytes. Do you really have more than 10 separate fields as the key, or are you confusing the number of columns with the number of keys?
_________________
Frank Yaeger - DFSORT Development Team (IBM)
Specialties: JOINKEYS, FINDREP, WHEN=GROUP, ICETOOL, Symbols, Migration
DFSORT is on the Web at:
www.ibm.com/storage/dfsort

shuko · Beginner Joined: 08 Nov 2005 Posts: 73 Topics: 20

Frank,

My requirement.

File1

kolusu · Posted: Tue Feb 21, 2006 2:52 pm Post subject:

Shuko,

Since the fields you are checking aren't conscuetive , you need to use an inrec/outrec and put all the fields you want to check at the end of every record and then use the select statement on the new padded fields.

let us say you have a 80 byte file and you want to check the 2 byte fileds at pos 20, 24, 28, ...

Frank Yaeger · Posted: Tue Feb 21, 2006 4:18 pm Post subject:

Shuko,

The way I approach this kind of problem is to splice each pair of records together and then use INCLUDE logic to determine which record goes to which file according to the "rules" you need. Here's a DFSORT job to do that based on your example and my interpretation of your stated rules. Although I got the output you asked for, I may not have the INCLUDE conditions exactly the way you want them for your rules, but you can adjust them as needed.

shuko · Beginner Joined: 08 Nov 2005 Posts: 73 Topics: 20

Frank,Kolusu

Thank you
I tried both the codes, Kolusu code in statement ON(81,40,CH) returned an error "FIELD BEYOND MAXIMUM RECORD LENGTH."
CTL2CNTL was declared in the JCL but not used, is there a statement missing?. Frank your code gave me the expected
result,except that I had the record with key 1111 in both the OUT files.

kolusu · Posted: Wed Feb 22, 2006 10:44 am Post subject:

Frank Yaeger · Posted: Wed Feb 22, 2006 1:12 pm Post subject:

shuko · Beginner Joined: 08 Nov 2005 Posts: 73 Topics: 20

Thank you Kolusu,Frank.

Frank,I had line numbers in my JCL from pos 73-80. I removed the line numbers and the code worked as expected. Thank you.

shuko · Beginner Joined: 08 Nov 2005 Posts: 73 Topics: 20

Frank,
I tried to implement your working example, for my actual situation.
LRECL=8000, 3 keys and check to be done from position 362 in both files.
I get the error ICE221A 1 INVALID FIELD OR CONSTANT IN OUT1 IFTHEN 0 CONDITION 2. Obviously I have an error in the second realational condition in OUT1 and cannot figure out the reason.

The code

kolusu · Posted: Thu Feb 23, 2006 5:40 am Post subject:

shuko,

Whenever you say something does not work you need to post the entire error messages.

Post your entire TOOLMSG and DFSMSG here

Kolusu
_________________
Kolusu
www.linkedin.com/in/kolusu

shuko · Beginner Joined: 08 Nov 2005 Posts: 73 Topics: 20

TOOLMSG

Frank Yaeger · Posted: Thu Feb 23, 2006 12:16 pm Post subject:

Shuko,

When you get a DFSORT error message, the first thing you should do is look it up. You can use LOOKAT to do that, or go directly to "z/OS DFSORT Messages, Codes and Diagnosis" which you can access online from:

www.ibm.com/servers/storage/support/software/sort/mvs/srtmpub.html

If you look up ICE221A, you'll see this:

One of the following errors was detected:
o the length for a field with a format other than SS was greater than 256

That's the reason for the ICE221A - you have 362,7638,CH,NE,8005,7638,CH so both lengths are 7638 which is greater than 256. You'll need to break this up into 256 byte segments. I'd suggest using DFSORT Symbols to make this easier. Add: