On Tue, 10 Nov 2015 19:47:49 -0500, Tony Harminc wrote:
>On 10 November 2015 at 19:04, Bharath Nunepalli wrote:
>> The confusing part here is the size of the Output dataset.
>> Input dataset is 1.7 GB, but the Output dataset is 8.1 GB.
>> Do you think this happened due to BLKSIZE(132) value??
>
>Yes, certainly. BLKSIZE of 132 is hopelessly space-wasting (and slow).
>If you need to preallocate, you can do that without specifying a
>BLKSIZE (SMS should choose a good one for you), or specify
>BLKSIZE=27998, which is half-track for a 3390.
>
Yes. With LRECL=128,BLKSIZE=132, you'll get only one record per
block. Most of your space will be interblock gaps.
But I'm skeptical that such an adverse input format could result in
even worse output. I'd like to see the output of this Rexx EXEC applied
to both your input and output data sets:
user@OS/390.24.00: cat rexxdump
/* Rexx */ signal on novalue; /*
************************************************
Doc: simple minded data set dump. Requires z/OS 2.1 or higher.
*/
if 0 then trace R
Dataset = userid()'.ISFUNLD.BADSYM.JOB08364.PAX.Z' /* Supply yours. */
say BPXWDYN( 'alloc rtddn(DD) dsn('Dataset') shr recfm(U) blksize(32760)
msg(2)' )
do I = 1 to 10
address 'MVS' 'EXECIO 1 DISKR' DD '(stem L.'
if RC<>0 then return( RC * (RC<>2) )
say 'Block'right( I, 3 ) 'is'right( length( L.1 ), 6 ) 'bytes.'
end I
/*
*****************************************************************************/
user@OS/390.24.00:
user@OS/390.24.00: rexxdump # sample output
0
Block 1 is 27998 bytes.
Block 2 is 4294 bytes.
user@OS/390.24.00:
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
