IBM Mainframe Forum

by **Steve Coalbran** » Fri May 09, 2014 3:38 pm

BillyBoyo wrote:...data.

Sorrry Billy - missed your post while I was reading my sudden influx of Notes tripe!

by **Steve Coalbran** » Fri May 09, 2014 3:51 pm

Sorry Billy, that just gives me...

ICE000I 1 - CONTROL STATEMENTS FOR 5694-A01, Z/OS DFSORT V1R12 - 12:12 ON FRI MA
OPTION COPY
OUTFIL FTOV,BUILD=(1,2,BI,TO=ZD,LENGTH=3,X,5,132)
ICE251A 9 MISSING RDW OR DATA FOR SORTOUT : REASON CODE 05, IFTHEN 0

I'm not sure that this is going to work.
I don't want the RDW, record-length or anything else in the output but (in REXX) SUBSTR(recordwithoutRDW,12)

by **BillyBoyo** » Fri May 09, 2014 4:11 pm

My sorry. I misspelled VTOF... VTOF is Variable To Fixed. FTOV is Fixed To Variable. Since you are starting with variable, and I was trying to show fixed, it should have been VTOF.

I wasn't suggesting it as what you want for your solution, but just so you could see your data with the lenghts as human-readable, so that you know what the blank lines actually look like. I suspect a blank like will show 005 and then a load of blanks, but most of those blanks are SDSF padding, not on the data. Sorry to confuse.

by **Steve Coalbran** » Fri May 09, 2014 5:58 pm

Thanks Billy,
Ha ha ha! Seems to be going backwards? :lol:

Although it didn't fail!?

Select all

SDSF OUTPUT DISPLAY K248610K JOB63408 DSID 105 LINE 0 C
COMMAND INPUT ===> SC
********************************* TOP OF DATA *******************
124 1 ISRSUPC - MVS/PDF FILE/LINE/WORD/BYTE/SFOR COMPARE UTI
081 LINE-Ä SOURCE SECTION SRCH DSN: SYS4.EQQ
005
099 467181 05/09 00.00.01 EQQE016I ETT HAS SUCCESSFULLY ADDED
059 467182 05/09 00.00.01 EQQE016I BS40P.OPC.START.MVS0
112 467183 05/09 00.00.01 EQQE040W JOB DSPS6571(JOB86033) CANN
112 467184 05/09 00.00.01 EQQE040W JOB DSPC4486(JOB86034) CANN
112 467185 05/09 00.00.01 EQQE040W JOB DSP06566(JOB86035) CANN
...

I was hoping my SYSOUT (for SORTOUT) looked more like...

Select all

SDSF OUTPUT DISPLAY K248610K JOB63408 DSID 105 LINE 0 C
COMMAND INPUT ===> SC
********************************* TOP OF DATA *******************
05/09 00.00.01 EQQE016I ETT HAS SUCCESSFULLY ADDED ENDEVORPROD5 F
05/09 00.00.01 EQQE016I BS40P.OPC.START.MVS0
05/09 00.00.01 EQQE040W JOB DSPS6571(JOB86033) CANNOT BE HANDLED.
05/09 00.00.01 EQQE040W JOB DSPC4486(JOB86034) CANNOT BE HANDLED.
05/09 00.00.01 EQQE040W JOB DSP06566(JOB86035) CANNOT BE HANDLED.
...

DFSORT seems to be very hard at handling null VB records ?
REXX solution is looking good?

by **NicC** » Fri May 09, 2014 6:19 pm

Hi Steve - this seems to be going nowhere! Can you post the latest SYSIN that you tried and the first few lines ofyour SORTIN in ISPF edit - not SDSF edit - with the third line HX'd (and do not allow edit to pad your short lines!). A 3.2 of the dataset may be useful for confirmation that we are all on the same song sheet!

by **BillyBoyo** » Fri May 09, 2014 6:50 pm

I think I'm just not being clear.

I suggested the short SORT so that we could see what the blank line looks like. And we can now confirm that it has a length of 005.

You seem to be favouring an F output, so we go back to Kolusu's suggestion:

Select all

OPTION COPY,VLSCMP

OMIT COND=(8,07,CH,EQ,C'ISRSUPC',OR,
7,06,CH,EQ,C'LINE-Ä',OR,
7,10,CH,EQ,C' ')

INREC OVERLAY=(134:X)
OUTFIL VTOF,BUILD=(16,122)

Knowing that the blank lines are only five bytes long:

Select all

OPTION COPY,VLSCMP

OMIT COND=(1,2,BI,EQ,5,OR,
8,07,CH,EQ,C'ISRSUPC',OR,
7,06,CH,EQ,C'LINE-Ä',OR,
7,10,CH,EQ,C' ')

INREC OVERLAY=(138:X)
OUTFIL VTOF,BUILD=(16,122)

If the 7,10 is solely for the blank lines, that can now be removed, but I'll leave it.

You have 137 bytes, so an adjustment to the OVERLAY, it puts a space at position 138, outside your data, automatically space-padding to that point.

Then the BUILD on OUTFIL can find data, and it will be pre-space-padded rather than binary-zero padded.

The OUTFIL will ignore the RDW (four bytes) and the first 11 of each record, and give you data from position 16 for a length of 122 bytes.

Then there's the stuff at the bottom which we've not seen yet. Just bear in mind that what look like trailing spaces in SDSF will not be there.

by **skolusu** » Fri May 09, 2014 9:48 pm

Steve Coalbran wrote:DFSORT seems to be very hard at handling null VB records ?

Steve,

Trust me DFSORT is more than capable of handling Null VB records. IMHO , I think we are just trying to tape the holes in the solution that you came up as we really don't know what the data looks like. If your intention is to simply pick the the data from byte 16 , then simply eliminate the records that have length less than that using OMIT condition.

It would be as simple as this

Select all

//SYSIN DD *
OPTION COPY,VLSCMP

OMIT COND=(1,02,BI,LT,16,OR,
8,07,CH,EQ,C'ISRSUPC',OR,
7,06,CH,EQ,C'LINE-Ä',OR,
7,10,CH,EQ,C' ')

INREC BUILD=(1,4,16)
//*

Steve Coalbran wrote:REXX solution is looking good?

Try running the rexx exec against a file with more than million records and see how much time it takes.

Btw if you need further help on this issue, please feel free to contact me on sametime or email me directly(available in bluepages) or even use DFSORT hotline dfsort@us.ibm.com

by **Steve Coalbran** » Sat May 10, 2014 4:57 pm

Thanks,
The format of the data is standard for ISRSUPC except that they have put it to a VB 137 dataset!
An excerpt of the data is included in the JCLstream included below.
Assume NULLS ON meaning that any blank lines are length 0.

Select all

EDIT K248610.USER.JCL(STEFANK) - 01.06 Columns 00001 00072
Command ===> Scroll ===> CSR
****** ***************************** Top of Data ******************************
000001 //K248610K JOB (BSJ0,P,B),'STEFANK',CLASS=B,MSGCLASS=O,NOTIFY=&SYSUID
000002 //*
000003 //*--------------------------------------------------------------------
000004 //* SORT EXTRACT FROM SUPERC
000005 //*--------------------------------------------------------------------
000006 //SORT EXEC PGM=SORT
000007 //SYSOUT DD SYSOUT=*
000008 //SORTIN DD DISP=SHR,DSN=K248610.MLOG.ERROR.COPY
000009 //SORTOUT DD SYSOUT=* ,OUTLIM=100
000010 //INPUTDS DD *
=COLS> ----+----1----+----2----+----3----+----4----+----5----+----6----+----7--
000011 BROWSE SYS4.EQQ.OPCA.MLOG.ERROR.COPY.G9909V00
000012 COMMAND ===>
000013 ********************************* TOP OF DATA ********
000014 1 ISRSUPC - MVS/PDF FILE/LINE/WORD/BYTE/SFOR COMP
000015 LINE-Ä SOURCE SECTION SRCH DSN:
000016
=NOTE= v- from here --->
000017 340948 05/08 00.00.01 EQQM923I JOBLOG FOR BWFXRATE
000018 340949 05/08 00.00.01 EQQM923I FOR OPERATION: BWFX
000019 340950 05/08 00.00.02 EQQE016I ETT HAS SUCCESSFULL
====== - - - - - - - - - - - - - - - 37746 Line
000020 502872 05/09 07.59.33 EQQE016I BJ80P.STARTA.PXSNIJ
000021 502873 05/09 07.59.49 EQQE016I ETT HAS SUCCESSFULL
000022 502874 05/09 07.59.49 EQQE016I PDGSX.XTCPNON.SRSTA
000023
000024 1 ISRSUPC - MVS/PDF FILE/LINE/WORD/BYTE/SFOR COMP
000025 SEARCH-FOR SUMMARY SECTION SRCH DSN:
000026
000027 LINES-FOUND LINES-PROC DATASET-W/LNS DATASET-WO/LN
000028 35694 562598 1 0
000029
000030 PROCESS OPTIONS USED: ANYC MIXED
000031
000032 THE FOLLOWING PROCESS STATEMENTS (USING COLUMNS 1:72)
000033 SRCHFOR '05/09'
000034 ISRS024W EXTRA DATA DETECTED AFTER NORMAL STATEMENT E
000035
000036 ISRS004I LISTING LINES MAY BE TRUNCATED DUE TO LIMITI
000037 //*---- latest try -------------------------------------------------------
000038 //SYSIN DD *
000039 OPTION COPY
000040 OUTFIL VTOF,BUILD=(1,2,BI,TO=ZD,LENGTH=3,X,5,132)
000041 //*
****** **************************** Bottom of Data ****************************

I think it unlikely that this report will hit 1M lines. However (despite being a complite Rexxist) I still think a DFSORT solution would be much faster and more maintainable.
The dsired output is erely all real data lines from pos-16, no line-numbers, headers, trailers, sumaries,...

(see the inserted NOTEline?)

by **BillyBoyo** » Sat May 10, 2014 5:14 pm

Steve, your latest try is still my control cards which were just to show what the actual record-lengths were alongside the data that you see. I think you've again missed out on reading some posts.

I suggest you try some of the more recent examples, which were designed to get what you had thought you had coded for originally.

by **Steve Coalbran** » Sat May 10, 2014 6:16 pm

OK I just wanted a working solution I guess. Like te REXX one I supplied.
None of the answers I have received have taken away the blank lines and I haven't even begun to address dropping the footers yet via DFSORT.

IBM Mainframe Forum

SORT VB records from SUPERC to remove headers & drop line#s

Re: SORT VB records from SUPERC to remove headers & drop lin

Re: SORT VB records from SUPERC to remove headers & drop lin

Re: SORT VB records from SUPERC to remove headers & drop lin

Re: SORT VB records from SUPERC to remove headers & drop lin

Re: SORT VB records from SUPERC to remove headers & drop lin

Re: SORT VB records from SUPERC to remove headers & drop lin

Re: SORT VB records from SUPERC to remove headers & drop lin

Re: SORT VB records from SUPERC to remove headers & drop lin

Re: SORT VB records from SUPERC to remove headers & drop lin

Re: SORT VB records from SUPERC to remove headers & drop lin