searching and updating particualr record

gopinathmk · New User Joined: 17 Nov 2006 Posts: 15

I have input file of LRECL=133 where I have got 65 records for every employee having diff information on each record.Out of 65th I need to check 7th record(or 7th line) 34th position of every employee and need to search whether I have same value in Second file .Second file has LRECL =80 but will contain only 4 digit value.If i found same value in 7th record of 1st file present in as that of second file I need to place N on 125th position in 7th record of the first file.I need to search second file because I dont know where the value will be may be at last

For Ex think following line is 7th record of 1st employee and have got 0830 value in 34th position now i need to search whether 0830 value present in second file.If yes I need to Place N in 125th column of same 7th record of an employee
34th Position 125th Position
11889 DOUGLAS DR 0830 0030 053 N

Please let me know code that can perform better?

Robert Sample · Posted: Mon Feb 16, 2009 11:41 pm

gopinathmk · New User Joined: 17 Nov 2006 Posts: 15

Robert,

Sorry I mean to say I want code in an effective way and I currently dont have any code.Please let me know your ideas on the same

ykishor · New User Joined: 11 Aug 2007 Posts: 24 Location: my pc

I guess your first file is a VSAM . If not create it that way because the performance wil be too good. Create the key in such a way that for a particular employee u can identify all the 65 records independently.

Using the key read the first file and then perform your required operation...
It would be good if you could specify the second file as VSAM too.

REMEMBER to open our ur first file in I-O mode...and take file backups before running the programs

gopinathmk · New User Joined: 17 Nov 2006 Posts: 15

No Its not a vsam File kishore and its very difficult to have it as VSAM since each record of an employee is not having anything unique in all records.....

ykishor · New User Joined: 11 Aug 2007 Posts: 24 Location: my pc

if that is the case how do u identify to which employee a particular record belongs if there is not anything common...

if its not a VSAM create a temp VSAM and load the flat file to that....

Robert Sample · Posted: Wed Feb 18, 2009 5:09 am

Read all 65 records into an array, then use reference modification to identify the appropriate fields to access and change. You can then write all 65 array elements to your output file. I don't think you can do this without having an input file and an output file unless you change the file from sequential to VSAM.

dick scherrer · Posted: Wed Feb 18, 2009 5:33 am

Hello,

ykishor · New User Joined: 11 Aug 2007 Posts: 24 Location: my pc

we can have a different approach.....using the VSAM itself...

1. Use VSAM in I/P mode
2. create the sequential O/P
3. REPRO to the VSAM(if required as ur i/p file is PS)

In this case the proceess is faster and VSAM rebuiliding is not required in case of an abend is not there.

gopinathmk · New User Joined: 17 Nov 2006 Posts: 15

I have input file of LRECL=133 where I have got 65 records for every employee having diff information on each record.Out of 65th I need to check 7th record(or 7th line) 34th position of every employee and need to search whether I have same value in Second file .Second file has LRECL =80 but will contain only 4 digit value.If i found same value in 7th record of 1st file present in as that of second file I need to place N on 125th position in 1st record of the first file.I need to search second file because I dont know where the value will be,it may be at last also.

Remember I need to update in First record based on values of 7th record and second file value

For Ex think following line is 7th record of 1st employee and have got 0830 value in 34th position now i need to search whether 0830 value present in second file.If yes I need to Place N in 125th column of 1st record of every employee(not in the same 7th records but 1st record)

11889 DOUGLAS DR 0830 0030 053 N

Robert Sample · Posted: Fri Feb 20, 2009 9:25 pm

Have you looked at this: www.ibmmainframes.com/viewtopic.php?t=38137&highlight=

gopinathmk · New User Joined: 17 Nov 2006 Posts: 15

Yes, But here requirement is changed and I want to update the 1st record and not 7th record.

Robert Sample · Posted: Fri Feb 20, 2009 11:49 pm

First record, 7th record, 65th record ... it doesn't matter. The same basic method is used. Read all records into an array, do what you want using the array, then output all the array elements to your file.

dick scherrer · Posted: Sat Feb 21, 2009 1:40 am

Hello,

Any more discussion should be posted n the original topic (link above).

Changing which record is not a requirement change, , , It is just a duplicate post.

d

gopinathmk · New User Joined: 17 Nov 2006 Posts: 15

It will be more helpfull if i can get syntax for the following.

I want to read 65 records of lrecl 150 each from a input flat file until employee number changes and need to store it in an array.after storing I need to retrieve piece of data from 7th record and need to search in a table if same value present in table also i need to retreive 1st record from array and need to update 125th position and need to write full 65 records to output file and need to proceed with second 65 records for an second employee like wise till the end of first input file

dick scherrer · Posted: Sat Feb 21, 2009 10:23 pm

Hello,

Robert Sample · Posted: Sat Feb 21, 2009 10:23 pm

gopinathmk · New User Joined: 17 Nov 2006 Posts: 15

Can DFSORT help me in doing following:

1) Input file with LRECL=140 will have 65 records for each employee and can have several employees like wise and each employee with 65 records

2)second input file with lrecl=80 but got data only in first 4 columns(i.e Pic X(4))

Now i need to search 4 columns from 34th position of 7th record out of 65 records of each employee and I need to search whether same 4 digit value present in second input file.If Yes I need to update 125th position of 1st record of 65 records of every employee and need to write all 65 records to output file if value doesnt found in input file i need to write whole 65 records without updating any record. Like wise this should continue till last employee in input file

Robert Sample · Posted: Sat Feb 21, 2009 10:32 pm

Let's see -- you didn't get somebody to write your program for you in the COBOL forum, so now you're going to try in the SORT forum ... wanna bet it doesn't work here, either?

CICS Guy · Posted: Sat Feb 21, 2009 10:39 pm

You have asked the same general question three times.
First as a Syncsort question that got moved to COBOL, second as a COBOL question and third as a DFSort question.
Wht kind of solution are you looking for?

ykishor · New User Joined: 11 Aug 2007 Posts: 24 Location: my pc

Dear All

Can we all be a little more polite with the language we use.....agreed that many a time a lot of us want to get the job done in an easy way....but in the same breath I wish to add that nobody has forced anyone of us to join the forum and help others....we all are here to help each other out when one get's stuck...

...and also we must continue to toil and ask for codes only when one needs a parachute.....self-learning is the best way to move forward....

...a note of thanks to all the people who are continuously lending a helping hand...

Cheers....

CICS Guy · Posted: Sat Feb 21, 2009 11:27 pm

dick scherrer · Posted: Sat Feb 21, 2009 11:36 pm

Hello,

It won't work here as i'm locking the topic and relocating it and will issue a formal warning if this shows up again. . . In any part of the forum. . .

If any amount of time had been spent working on the reuirement rather than trying to get others to do the work, it would be done now. It has only been 5 days or so. . .

ykishor · New User Joined: 11 Aug 2007 Posts: 24 Location: my pc

to CICS Guy:

Robert Sample · Posted: Sun Feb 22, 2009 12:14 am

Raw code is about 200 lines of COBOL without error checking (I wrote the program this afternoon to verify that). Surely even a complete COBOL novice can write a couple hundred lines of code in 5 days -- allowing for some testing as well -- which means the code could have been completed by now if the original poster had just coded up the program instead of trying for days to get somebody to do it.