I want all the 11111 records to go to file 1. 22221 records to file 2 etc...I am not sure of all the combinations that can occur in the positions 1: to 5:.
Joined: 02 Dec 2002 Posts: 1618 Topics: 31 Location: San Jose
Posted: Mon Jul 28, 2003 7:44 pm Post subject:
Kolusu,
Since the field has 5 bytes and the constant has 5 bytes, you INCLUDE parameter of:
INCLUDE=(1,5,SS,EQ,C'11111')
is equivalent to:
INCLUDE=(1,5,CH,EQ,C'11111')
Not sure what you're trying to do here, but your INCLUDE parameters get you the '11111' records in file1, the '44444' records in file4 and the '55555' records in file5. It doesn't get the '22221' records in file2 or the '33322' records in file3.
Anyway, I think the idea is to switch to a new file every time the key changes. I don't know of any reasonable way to do that unless we know the actual values of the keys (which krk123 said wasn't the case), and we can hardcode the DDs (which krk123 may or may not want). _________________ 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
Thanks a lot Kolusu and Frank. As I dont know the different combinations that can occur in the key field, I dont think there is any other way of doing this.
Hi Kolusu, Frank,
what is the maximum outfiles we can use. I.e how many sortof files we can use...(SORTOF01, sortof02 ...how far we can go)
Is there a limit?
Thanks,
krk
Joined: 02 Dec 2002 Posts: 1618 Topics: 31 Location: San Jose
Posted: Tue Jul 29, 2003 11:05 am Post subject:
For DFSORT, there's no practical limit to the number of OUTFIL data sets you can use as long as you give DFSORT enough storage for the buffers. I just did a test using 1000 DDs and OUTFIL statements as shown below, with REGION=256M, and it ran successfully (I didn't try more than 1000, so I don't know what the actual limit is):
_________________ 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
Joined: 26 Nov 2002 Posts: 12378 Topics: 75 Location: San Jose
Posted: Tue Jul 29, 2003 2:37 pm Post subject:
Krk123,
sorry about my previous post. I over looked your input and suggested a solution which frank was kind enough to correct me. Thanks frank.I was in a hurry when I posted that solution. Here is another approach for your problem.
We first take the input file and sum sort it on the first 5 bytes to get the unique keys. From these unique keys we write dynamic control cards and file allocations and submit a job to the INTRDR.I assumed that you will have a max of 500 unique keys at any give time.If you have more than 500 , they will all be stored in a seperate dataset which you can process as if it is the input file.
Similary files DD2 thru DD5 will have file allocations for files101 thru file500
Step0200 in the above jcl will merge all these control card and dynamic allocations files together and submits a job to the INTRDR which will actually split the records into various files.
Joined: 02 Dec 2002 Posts: 1618 Topics: 31 Location: San Jose
Posted: Tue Jul 29, 2003 3:50 pm Post subject:
Kolusu,
Good idea. But I'm curious why you split the generated control statements into five files instead of just putting them all in one file? Likewise for the generated DD statements? _________________ 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
Joined: 26 Nov 2002 Posts: 12378 Topics: 75 Location: San Jose
Posted: Tue Jul 29, 2003 5:12 pm Post subject:
Frank,
Good catch! I was just afraid of "excess of control cards" error. I also was thinking about the max no: of DD names in a jcl. So splitting into 5 or 6 files can be easy as they can be used to submit multiple jobs at the same time. Ideally I would have to create 5 control cards datasets with 200 control statements and submit 5 jobs to the intrdr to execute simultaneously there by creating 1000 datasets in one go. If krk123 is willing to pursue this solution then I will post the updated solution.
Joined: 02 Dec 2002 Posts: 1618 Topics: 31 Location: San Jose
Posted: Tue Jul 29, 2003 5:49 pm Post subject:
Kolusu,
I guess the "excess of control cards" error is a Syncsort problem that DFSORT doesn't have. I don't know what the max number of DD names in a JCL is. As I said, I had 1000 DDs in the DFSORT job I tested and that didn't cause a problem, but I didn't run it through the internal reader (maybe that makes a difference?). Anyway, just curious. _________________ 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
Hi Kolusu & Frank,
here is the version I used. It worked great. To be frank initially it was above my head when kolusu posted. I did a lot of experimenting and reading to understand this. I still have some doubts, but not sure whether I can ask these here as they are kind of basic doubts..
Any ways, thanks a lot for kolusu and frank. I really appreciate your time and help on this.
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum