View previous topic :: View next topic |
Author |
Message |
ronderstal Beginner
Joined: 25 Apr 2007 Posts: 5 Topics: 2
|
Posted: Wed Aug 24, 2011 8:26 pm Post subject: Find multiple relations between records |
|
|
I am looking for a solution other then REXX and hope that Sort, ICEtool can offer me this. the REXX that is used now is using a lot of cpu.
We runnig daily jobs, were we want to find (extended predecessor) relations between jobs.
We want only the relations of the jobs that are specified. these jobs are
in a file1. dcb file 1 Recfm=fb, lrecl=132.
The records in file1 must be compared to file2.
record file1= jobname daynr time ws_oper , see picture 1
records File2= jobname daynr time ws_oper jobname daynr time ws_oper
see picture 2, it looks like file1
I not have any acces to a public server, so i can not post my Images.
When needed I can mail them.
dcb file2 Recfm=fb, lrecl=150.
File2 is giving the Predecessor relations were the right part is (kolom 2)
is de the predecessor of kolom1
What is done
I am looking for the relations of first record of file1 between the record relation given in File2
when looking for relations of first record in file1 in file2
file1 Code: |
AA100100 24 12.00 PR1_020
|
file2 relation found Code: |
AA100100 24 12.00 PR1_020 XQGHUNL 24 12.00 PR_020
AA100100 24 12.00 PR1_020 YYFP100 24 12.00 CIMS_005
|
so it founds the first 2 predecessors Code: |
XQGHUNL 24 12.00 PR_020
YYFP100 24 12.00 CIMS_005 |
these must be written to a d.s and should look something like Code: |
AA100100 24 12.00 PR1_020 XQGHUNL 24 12.00 PR_020
--------------------------------- YYFP100 24 12.00 CIMS_005 |
then it must look for the predecessors of these two jobs prev.found.
so looking for XQGHUNL 24 12.00 PR_020 it will find: Code: |
XQGHUNL 24 12.00 PR_020 XQGHUNL 24 12.00 CDB2_010
these must be written and be ADDED to previoius d.s
--------------------------------- XQGHUNL 24 12.00 CDB2_010
|
then it must look for the predecessors of these two jobs prev.found. Code: |
so looking for YYFP100 24 12.00 CIMS_005 it will find predecessor::
YYFP100 24 12.00 CIMS_005 YYRC700A 24 12.00 CIMS_055
YYFP100 24 12.00 CIMS_005 YYRC950 24 12.00 CIMS_025 |
these PRED. RELATIONS must be written and added to prev. ds.
and the found pred relations also have to be looked for, and so on.
until find no more then for the next record of f ile1 has to be done the same
evntually it will look like this for record1 of file1 Code: |
AA100100 24 12.00 PR1_020 XQGHUNL 24 12.00 PR_020
--------------------------------- YYFP100 24 12.00 CIMS_005
--------------------------------- XQGHUNL 24 12.00 CDB2_010
--------------------------------- YYRC700A 24 12.00 CIMS_055
--------------------------------- YYRC950 24 12.00 CIMS_025
|
I added the ---- for outlinig purposes
Where in this case AA100100 act as a label or start of a group.
the number of records for file1 can be up to 1000.
the number of records for file2 can be up 311000 and more.
Realtiosn can be to a few hundred and moer deep
The output dataset can be as wide as needed about lrecl 27000
I hope there is a solution for this in SORT, ICETOOL |
|
Back to top |
|
 |
kolusu Site Admin

Joined: 26 Nov 2002 Posts: 12378 Topics: 75 Location: San Jose
|
Posted: Wed Aug 24, 2011 9:54 pm Post subject: |
|
|
ronderstal,
I am kind of lost with the whole description. How about showing 5-10 records from input file 1 and then pick all the relations from file 2 and show them here. After that show us the output from those records so that we get a clear picture.
Also I added code tags to your posts that it is easy to align the data . Look at this topic for example of adding code tags
http://www.mvsforums.com/helpboards/viewtopic.php?p=19031#19031 _________________ Kolusu
www.linkedin.com/in/kolusu |
|
Back to top |
|
 |
ronderstal Beginner
Joined: 25 Apr 2007 Posts: 5 Topics: 2
|
Posted: Thu Aug 25, 2011 2:17 pm Post subject: |
|
|
Try a test. BBcode.
Code: |
***************************** Top of Data ***
test
|
ahhh nice, so easy
Not have the request details at hand, have to get them from work.
Send )part of) the req. by email.
Ronderstal |
|
Back to top |
|
 |
|
|