|
View previous topic :: View next topic
|
| Author |
Message |
vasanthz
Global Moderator
Joined: 28 Aug 2007
Posts: 1742
Location: Tirupur, India
|
|
|
|
Hi,
I have a dataset which looks like,
| Code: |
DSN VOLSER NEXTVOL
AAAA V0002 V0003
AAAA V0001 V0002
AAAA V0004
AAAA V0003 V0004
BBBB A0001
BBBB A0004 A0003
BBBB A0003 A0002
BBBB A0002 A0001
BBBB A0005 A0004
CCCC K0001
DDDD J000B J000C
DDDD J000C
DDDD J000Z J000B |
The requirement is to create a column VOLSEQ which is volume sequence number.
Output required:
| Code: |
DSN VOLSER NEXTVOL VOLSEQ
AAAA V0001 V0002 1
AAAA V0002 V0003 2
AAAA V0003 V0004 3
AAAA V0004 4
BBBB A0005 A0004 1
BBBB A0004 A0003 2
BBBB A0003 A0002 3
BBBB A0002 A0001 4
BBBB A0001 5
CCCC K0001 1
DDDD J000Z J000B 1
DDDD J000B J000C 2
DDDD J000C 3 |
The logic is,
For DSN = AAA, VOLSER V0004 does not have NEXTVOL, so it is the last VOLSER in the sequence(VOLSEQ=4).
Next, the record
AAAA V0003 V0004
has V0004 as NEXTVOL, so it is the last but first VOLSER in sequence(VOLSEQ=3).
Next, the record
AAAA V0002 V0003
has V0003 as NEXTVOL, so its VOLSEQ=2 and so on..
Could you please let me know how this could be done with DFSORT/REXX/SAS?
The input and output file's 1st record is just for understanding and can be ignored in actual job. the input and output files can be assumed as FB LRECL 80.
Thanks & Regards, |
|
| Back to top |
|
 |
Bill Woodger
Moderator Emeritus
Joined: 09 Mar 2011
Posts: 7309
Location: Inside the Matrix
|
|
|
|
| What about WHEN=GROUP with KEYBEGIN on the DSN and PUSH the SEQ? |
|
| Back to top |
|
|
dick scherrer
Moderator Emeritus
Joined: 23 Nov 2006
Posts: 19244
Location: Inside the Matrix
|
|
|
|
Hi Vasanth,
As Bill mentioned, WHEN=GROUP can be your friend.
Suggest that you temporarily replace the "end of group" with all 9's so they will sort to the bottom rather than the top. . . When the output is generated, the blanks can be put back in (or the 9's left to indicate the end). |
|
| Back to top |
|
|
Bill Woodger
Moderator Emeritus
Joined: 09 Mar 2011
Posts: 7309
Location: Inside the Matrix
|
|
|
|
There shouldn't be a need to sort the file. The contents of the columns for the volumes is not relevant to the sequence if just OPTION COPY/SORT FIELDS=COPY and simply do an increasing sequence for the records in the group (same DSN).Sortinng on the volumes would loose/make difficult the actual sequence.
I suspect Vasanthz you got too close to the problem through knowing the "next" volume...
Or I'm missing something obvious... |
|
| Back to top |
|
|
vasanthz
Global Moderator
Joined: 28 Aug 2007
Posts: 1742
Location: Tirupur, India
|
|
|
|
Hi Bill and D,
Thanks for your time in responding.
| Quote: |
| Sortinng on the volumes would loose/make difficult the actual sequence. |
The VOLSER AND NEXTVOL can be in any order, they cannot be sorted.
Example for input.
| Code: |
DDDD J1111 J3333
DDDD J3333
DDDD J2222 J1111 |
the output has to be,
| Code: |
DDDD J2222 J1111 1
DDDD J1111 J3333 2
DDDD J3333 3 |
The source of the records is a snapshot file, which has 1 single record for each volume in which the dataset resides. So the volume names can be arbitrary and cannot be guaranteed to be in sequence.
Hope I clarified things.
Regards, |
|
| Back to top |
|
|
Skolusu
Senior Member
Joined: 07 Dec 2007
Posts: 2205
Location: San Jose
|
|
|
|
| vasanthz wrote: |
The VOLSER AND NEXTVOL can be in any order, they cannot be sorted.
Example for input.
| Code: |
DDDD J1111 J3333
DDDD J3333
DDDD J2222 J1111 |
the output has to be,
| Code: |
DDDD J2222 J1111 1
DDDD J1111 J3333 2
DDDD J3333 3 |
|
Without SORTING , how do you plan to arrange the records as if they are sorted? |
|
| Back to top |
|
|
vasanthz
Global Moderator
Joined: 28 Aug 2007
Posts: 1742
Location: Tirupur, India
|
|
|
|
Hi Kolusu,
Thanks for your time.
For the set of input records,
| Code: |
DDDD J1111 J3333
DDDD J3333
DDDD J2222 J1111 |
the record
DDDD J3333
does not have NEXTVOL column, so it is has to be the last in the group with VOLSEQ = 3, making required output record as
DDDD J3333 3
Next, for the input record
DDDD J1111 J3333
has J3333(VOLSER from previous record) as NEXTVOL so it has to be VOLSEQ = 2, making the required output record as
DDDD J1111 J3333 2
Next, for the input record
DDDD J2222 J1111
has J1111(VOLSER from previous record) as NEXTVOL, so it has to be
VOLSEQ = 1, so output record is
DDDD J2222 J1111 1
Hope I am not complicating this too much :S
P.S tried using the code tags for records, they look more confusing than as it is now (:
Regards, |
|
| Back to top |
|
|
Skolusu
Senior Member
Joined: 07 Dec 2007
Posts: 2205
Location: San Jose
|
|
|
|
vasantz,
If your intention is to make the Space record the last in the group and assign seq numbers then it is very easy.
| Code: |
//STEP0100 EXEC PGM=SORT
//SYSOUT DD SYSOUT=*
//SORTIN DD *
AAAA V0002 V0003
AAAA V0001 V0002
AAAA V0004
AAAA V0003 V0004
BBBB A0001
BBBB A0004 A0003
BBBB A0003 A0002
BBBB A0002 A0001
BBBB A0005 A0004
CCCC K0001
DDDD J000B J000C
DDDD J000C
DDDD J000Z J000B
//SORTOUT DD SYSOUT=*
//SYSIN DD *
INREC IFTHEN=(WHEN=INIT,OVERLAY=(81:C'0')),
IFTHEN=(WHEN=(20,5,CH,EQ,C' '),OVERLAY=(81:C'1'))
SORT FIELDS=(01,5,CH,A,
81,1,CH,A),EQUALS
OUTREC IFOUTLEN=80,
IFTHEN=(WHEN=GROUP,KEYBEGIN=(1,5),PUSH=(30:SEQ=1))
//* |
This will produce
| Code: |
AAAA V0002 V0003 1
AAAA V0001 V0002 2
AAAA V0003 V0004 3
AAAA V0004 4
BBBB A0004 A0003 1
BBBB A0003 A0002 2
BBBB A0002 A0001 3
BBBB A0005 A0004 4
BBBB A0001 5
CCCC K0001 1
DDDD J000B J000C 1
DDDD J000Z J000B 2
DDDD J000C 3
|
|
|
| Back to top |
|
|
vasanthz
Global Moderator
Joined: 28 Aug 2007
Posts: 1742
Location: Tirupur, India
|
|
|
|
Hi Kolusu,
Thanks for the code, but the output records has to be arranged as shown in the attachment.
Hope this makes things clear. |
|
| Back to top |
|
|
Skolusu
Senior Member
Joined: 07 Dec 2007
Posts: 2205
Location: San Jose
|
|
|
|
vasanthz,
I guess you are trying to find a relationship chain. I don't think sort would be an ideal choice here. I suggest you code a program for it. |
|
| Back to top |
|
|
vasanthz
Global Moderator
Joined: 28 Aug 2007
Posts: 1742
Location: Tirupur, India
|
|
|
|
| Kolusu, Thanks again for your thoughts. |
|
| Back to top |
|
|
Bill Woodger
Moderator Emeritus
Joined: 09 Mar 2011
Posts: 7309
Location: Inside the Matrix
|
|
|
|
You need to store all your information for one DSN in a table/array. You need to locate the entry fpr which the "next vol" is blank. Assign this the final sequence number. Take the volume from that entry and search for it as "next vol". Reduce the sequence number. Repeat until sequence equals one. Write table/array out. Repeat until end-of-file. Sort file on DSN/sequence.
Apologies for only reading what I thought I saw. |
|
| Back to top |
|
|
vasanthz
Global Moderator
Joined: 28 Aug 2007
Posts: 1742
Location: Tirupur, India
|
|
|
|
Bill, I did write a program with the logic you mentioned(without the array though),
ibmmainframes.com/viewtopic.php?t=58798&highlight=
but the program was making more than 4 passes of the same file.
The array idea you mentioned might save some passes of the file and make the program little more efficient. I will try out the array logic.
Regards, |
|
| Back to top |
|
|
expat
Global Moderator
Joined: 14 Mar 2007
Posts: 8797
Location: Welsh Wales
|
|
|
|
Vasanthz, take a look at the IGGCSIRX code that comes in SYS1.SAMPLIB
It lists volsers in order, and from that you can then arrange the table as required |
|
| Back to top |
|
|
vasanthz
Global Moderator
Joined: 28 Aug 2007
Posts: 1742
Location: Tirupur, India
|
|
|
|
Expat, the data is passed to us with many other DSN details in a single file from client LPAR(snapshot from some utility). So I will not be able to change it.
Regards, |
|
| Back to top |
|
|
|
|
