MVSFORUMS.com

[da_one]

Hi,I am facing problems trying to format the Trailer1 fields.

The control card
SORT FIELDS=(some criteria)
INCLUDE=( some conditions),
TRAILER1=(<part1>,<part2>,<part3>,COUNT,<part4>)

The requirement that I have is
<part1> = This need to be a 16 byte field, should contain the "High-value"
<part2> = 6 byte field, simple constant X(6), For eg. "ENDREC"
<part3> = 3 byte numeric field, but again a constant 9(3), For eg "008"
COUNT = This is the count function of the sort, but I need it to be in S9(09) COMP-3 format. I do not know what is the default format of the COUNT function.
<part5> = FILLER of 10 bytes

It would be great if anyone can help me with this.

[Frank Yaeger]

For information on how DFSORT formats relate to COBOL formats, see the "What are the equivalent DFSORT formats for various COBOL data types?" Ask Professor Sort item at:

http://www.ibm.com/servers/storage/support/software/sort/mvs/professor_sort/

It's not clear to me what you want:

<part1> - what do you mean by "High-value"? Do you mean a constant like X'FF.FFF' or C'9...9' or something else?
<part2> - just use C'ENDREC'
<part3> - what is the position, length and format of this numeric field? Do you want to convert it to a value like '008' or is it already a value like '008'?
COUNT - with DFSORT, you can format the COUNT to various printable formats in TRAILER1 but you can't directly convert it to PD format in TRAILER1. You could use TRAILER1 to get the COUNT as a ZD value using COUNT=(M11,LENGTH=9) and then use another COPY step to convert the COUNT to PD using p,9,ZD,TO=PD,LENGTH=5.
<part5> - if you want 10 bytes of blanks as the FILLER,you can use 10X. If you want something else as the FILLER, what do you want?

I can show you how to do what you want with DFSORT, but I need the answers to the questions above. Also, what is the RECFM and LRECL of the input data set?
_________________
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

[da_one]

Hi Frank,
<Part 1>: I need High-values as decribed by Cobol "High-values", in the first 16 bytes. I tried giving X'FFFFFFFFFFFFFFFF' earlier but it showing me some syntax errors. I think I am not giving it properly.
<Part 2> this follows Part 1 immediately
<Part 3> is 3 byte numeric field imediately after the above 2 fields. i.e 20-22 bytes.
<COUNT> Will I need another pass on my data to just reformat the COUNT field to comp-3 format ?
<Part 4> I just want the 10X filler.

It would be great if you could let me know. Thanks for such prompt response

[Frank Yaeger]

Well, you didn't really answer all of my questions, but I'll guess at what you need. I'm assuming that your input file has RECFM=FB and LRECL=80. I'm also assuming that you want your output data set to be FB, not FBA, without carriage control characters. Here's a DFSORT/ICETOOL job that will do what I think you want:

[da_one]

Hi Frank,
Yes, the file is FB type and u can assume 80 byte length. I want to format the TRAILER1 itself (Not the OUTREC) so that I get the High-values i.e 'FF' in the first 16 bytes. Is this possible. Currently when I give
TRAILER1=(16X'FF',C'ENDFIL',C'008',COUNT,10X)
it gives me the syntax error
<<TRAILER1=16X'FF',C'ENDFIL',C'008',COUNT,10X)
$
ICE007A 6 SYNTAX ERROR
>>
Frank, Also I want to avoid an extra pass over the data to format the count. If that cannot be achived in the single pass, I am comfortable in retaining the Count in ZD format, but in 9 bytes. But I have to get the High-Values in the first 16 bytes.

thanks for your prompt responses

[Frank Yaeger]

da one,

TRAILER1 does not support nX'hh', so you can't use TRAILER1 for 16X'FF'.

I assumed that you wanted to write the data records and a trailer record. Is that correct, or do you just want to write the trailer record (no data records)? Assuming you do want the data records:

Please try the job I gave you. It will give you what you want. It does two passes over the records, but the second pass is only for 1 record, so it's so close to a single pass that it doesn't matter.

To explain further:

In the first pass, I copy all of the records and create a "count" record with the ZD COUNT. In the second pass, I only read that one "count" record and reformat it to one "trailer" record in the form you want, which I MOD on to the data records.

So in the first pass, I'm reading n records and writing n+1 records. In the second pass, I'm reading 1 record and writing 1 record. Thus, in total, the job reads n+1 records and writes n+2 records. Using TRAILER1 (if you could), you would read n records and write n+1 records. The difference in the number of records read and written is insignificant.
_________________
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

[kolusu]

Da_one,

You are missing an opening parenthesis on your trailer statement.you should have

TRAILER1=(16X'FF',C'OUTREC',C,'008',COUNT,10X,80:X)

You cannot format the count to packed decimal format on the trailer parm. But as you said you are ok with the ZD format with 9 byte length

TRAILER1=(16X'FF',C'OUTREC',C,'008',COUNT=(M11,LENGTH=9),10X,80:X

The following JCl will give the desired results.

[Frank Yaeger]

Kolusu,

I don't understand what you're doing here. As I said, 16X'FF' is invalid in TRAILER1. You seem to be using 81,16 to try to get 16X'FF', but 81-96 doesn't contain 16X'FF'. I don't see any way to put 16X'FF' in 81-96 so TRAILER1 will pick it up and then remove it from 81-96 in ONE pass. What did you have in mind? You might want to try it to see if it works or not.

You also have C, '008' which is invalid syntax. It should be C'008'.
_________________
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

[kolusu]

Frank,

I actually had an INREC statement which I missed out on the posting.The actual control cards should be as follows

[Frank Yaeger]

Kolusu,

Oh, I see. This works because TRAILER1 picks up the 16X'FF' from 81-96 of the last INREC RECORD before the OUTREC parameter removes 81-96. Good exploitation of a subtle point.

BTW, you don't need the 80:X in TRAILER1 - the length of TRAILER1 will be set from the OUTREC parameter.

Of course, this doesn't produce the PD count field, but then I guess that was optional.
_________________
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

[Frank Yaeger]

With z/OS DFSORT V1R5 PTF UQ95214 or DFSORT R14 PTF UQ95213 (Dec, 2004), you can now use nX'hh' and stats=(to) in the TRAILERx parameter, so this can be done more easily and efficiently in one pass like this:

[s_shivaraj]

Hi,
When i tried the following the sort card, i am getting a syntax error , it will be great if some one can help me in that..

[Frank Yaeger]

You didn't show the error message, but I'm guessing from the * under the = that it's a WER message which means you're using Syncsort, not DFSORT. DFSORT supports COUNT=(edit). The error message with the * under the = is telling you that Syncsort does NOT support it. Note that DFSORT also supports COUNT=(to), e.g. COUNT=(TO=ZD,LENGTH=5).
_________________
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

[kolusu]

Shivraj,

As frank mentioned older version of syncsort does not support the formatting of the count field. However the new syncsort z/os 1.2 version supports the formatting. Here is an alternative way of getting the desired results.

[s_shivaraj]

Frank Yaeger, kolusu,

Its working,,, Thanks a lot
_________________
Cheers
Sivaraj S

'Technical Skill is the Master of complexity, while Creativity is the Master of Simplicity'