IBM Mainframe Forum

I am pretty new to DFSORT and have a REXX code which is currently doing

I have file 1 which has a list of datasets in this format :

1-3 6-50 (max dataset length 44)
xxx somedatasetname-1-hereupto44chars
xxx somedatasetname-4-hereupto44chars
xxx somedatasetname-7-hereupto44chars

and another file 2 also containing a list of datasets along with some other fields:

1-6 10-54 (upto 44 chars) 57-66 70-79 83-87 92-93
vvvvvv somedatasetname-1-hereupto44chars dd/mm/yyyy dd/mm/yyyy xxxxx zz
vvvvvv somedatasetname-2-hereupto44chars dd/mm/yyyy dd/mm/yyyy xxxxx zz
vvvvvv somedatasetname-3-hereupto44chars dd/mm/yyyy dd/mm/yyyy xxxxx zz
vvvvvv somedatasetname-4-hereupto44chars dd/mm/yyyy dd/mm/yyyy xxxxx zz
vvvvvv somedatasetname-5-hereupto44chars dd/mm/yyyy dd/mm/yyyy xxxxx zz
vvvvvv somedatasetname-6-hereupto44chars dd/mm/yyyy dd/mm/yyyy xxxxx zz
vvvvvv somedatasetname-7-hereupto44chars dd/mm/yyyy dd/mm/yyyy xxxxx zz


I want to compare these files and extract records from file 2 into a third file which has the dataset names as listed in file 1.

Outputfile:

vvvvvv somedatasetname-1-hereupto44chars dd/mm/yyyy dd/mm/yyyy xxxxx zz
vvvvvv somedatasetname-4-hereupto44chars dd/mm/yyyy dd/mm/yyyy xxxxx zz
vvvvvv somedatasetname-7-hereupto44chars dd/mm/yyyy dd/mm/yyyy xxxxx zz

I am currently using a REXX to do this, but that takes up quite a lot of time comparing each record one by one with the second field in file 2. I am new to DFSORT and want to see if DF/SORT can do it more quickly for me. Any ideas?

If you look at the document "DFSort - tricks.pds" and look at the trick "Pull records from a Master file" that would appear to do what you want. I cannot get onto my mainframe to code it up to test it. The document can be found here

http://www.ibm.com/support/docview.wss? ... g3T7000094

Thanks for pointing that out, NicC. I have a copy of that PDF and I checked the trick. The trick seems to do pretty much what I want to accomplish. However, there is an important difference. The key (account no as in the example trick) is variable length in my case. It is a dataset name which can be of any length (upto 44 chars).

Sorry, but I am pretty new to DF/SORT - can you advise what changes would be required to accomodate this requirement into the job?

Sorry, no I cannot. You will have to wait for someone who can!

MFsys,

The following DFSORT JCL will give you the desired results

Thanks, skolusu. Unfortunately, we are pretty behind on our DF/SORT maintenance (running April, 2006 DFSORT functions) and JOINKEYS functionality isn't available at our site . I'll look into the possibility of getting upto date on our DF/SORT maintenance.

MFsys wrote:Thanks, skolusu. Unfortunately, we are pretty behind on our DF/SORT maintenance (running April, 2006 DFSORT functions) and JOINKEYS functionality isn't available at our site . I'll look into the possibility of getting upto date on our DF/SORT maintenance.

Use the following DFSORT/ICETOOL JCL which will give you the desired results. I assumed that you don't have duplicates on the DSN name

That did the trick . Thanks skolusu.

I have been trying my luck with the solution above, and have been reading and learning about the DFSORT/ICETOOL instructions given here. However, there seems to be some problem here. Obviously, there might be some facts which I may have missed while asking the question.

File 1 has a total of 2505 records. To my knowledge these are not duplicate. However, when I ran a DFSORT job with following control cards I am left with 2438 records.

SORT FIELDS=(1,50,CH,A)
SUM FIELDS=NONE

So, I am concluding that there were probably some records which were duplicates and were removed with the above job.

I did the same with File 2, but the no. of records remain the same. Hence, I am presuming that there are no duplicates here. I now have 45706 records in File 2.



Now, I expect the REPORT to contain a max of 2438 records. However, I end up with 2708 records. This doesn't seem to be correct?? I don't want to / CAN'T check each record individually to find out which are the extra records that have been included here and why. What am I missing here?

Record format of my FILE1 dataset :

3 char field1
2 blanks
44 char field for datasetname



Record format of my FILE2 dataset :

3 blanks
6 char field1
3 blanks
44 char field for dataset name (to be compared with FILE1)
3 blanks
10 chars for field3
3 blanks
10 chars for field4
3 blanks
5 char for field5
4 blanks
2 char for field6

Ideally I would want the output to be something like this:

3 char from field1 of FILE1
3 blanks
44 char field for datasetname
3 blanks
6 char field1 from FILE2
3 blanks
10 chars for field3
3 blanks
10 chars for field4
3 blanks
5 char for field5
4 blanks
2 char for field6