Portal | Manuals | References | Downloads | Info | Programs | JCLs | Master the Mainframes
IBM Mainframe Computers Forums Index
 
Register
 
IBM Mainframe Computers Forums Index Mainframe: Search IBM Mainframe Forum: FAQ Memberlist Usergroups Profile Log in to check your private messages Log in
 

 

Get Record count in summary record for each group of records
Goto page 1, 2  Next
 
Post new topic   Reply to topic    IBMMAINFRAMES.com Support Forums -> DFSORT/ICETOOL
View previous topic :: :: View next topic  
Author Message
Atul Banke

New User


Joined: 08 Aug 2011
Posts: 5
Location: India

PostPosted: Fri Sep 23, 2016 4:17 pm    Post subject: Get Record count in summary record for each group of records
Reply with quote

Hi,

I need the get the record count of specific type of records and write that count to a summary record. There is one summary record for each group of detail records. E.g my input file format is as below

Code:
D|COUNT|111111111111|123.23|XYZ COMMERCE|NEW YORK CITY
R|11111110444A|15022016|0.34|XYZ AFFILIATE|BOSTON
R|11111110444A|15032016|2.44|XYZ AFFILIATE|BOSTON
R|11111110444B|15022016|12.76|XYZ BROTHERS|LA CITY
R|11111110444B|15032016|112.76|XYZ BROTHERS|LA CITY
D|COUNT|222222211111|73.23|ABC CONSTRUCTION|NEW YORK CITY
R|22222220111A|15022016|0.34|ABC CONSTRUCTION|BOSTON
R|22222220111B|15032016|71.63|ABC CONS.|BOSTON


Output Should have:
Code:
D|4|111111111111|123.23|XYZ COMMERCE|NEW YORK CITY
R|11111110444A|15022016|0.34|XYZ AFFILIATE|BOSTON
R|11111110444A|15032016|2.44|XYZ AFFILIATE|BOSTON
R|11111110444B|15022016|12.76|XYZ BROTHERS|LA CITY
R|11111110444B|15032016|112.76|XYZ BROTHERS|LA CITY
D|2|222222211111|73.23|ABC CONSTRUCTION|NEW YORK CITY
R|22222220111A|15022016|0.34|ABC CONSTRUCTION|BOSTON
R|22222220111B|15032016|71.63|ABC CONS.|BOSTON

That I just need the count of "R" records under each "D" type records on second field in pipe delimited record.

Could you please suggest a simplest way using DFSORT?

Thanks,
Atul

CODE'd
Back to top
View user's profile Send private message

Nic Clouston

Global Moderator


Joined: 10 May 2007
Posts: 1715
Location: UK

PostPosted: Fri Sep 23, 2016 4:25 pm    Post subject:
Reply with quote

What have you tried? Have you looked up any of the DFSort reference material - programmers guide, getting started, tricks?
Back to top
View user's profile Send private message
Arun Raj

Moderator


Joined: 17 Oct 2006
Posts: 2163
Location: @my desk

PostPosted: Fri Sep 23, 2016 11:53 pm    Post subject:
Reply with quote

Atul Banke,

The below DFSORT should give you something to start with. I have assumed the input to be FB with an LRECL=80. You might want to modify it as per your requirements.
Code:
//STEP0100 EXEC PGM=SORT                                         
//SYSOUT   DD SYSOUT=*                                           
//SORTIN   DD DISP=SHR,DSN ===>  Input Data set (FB/LRECL=80)     
//T1       DD DSN=&T1,DISP=(,PASS),UNIT=SYSDA                     
//T2       DD DSN=&T2,DISP=(,PASS),UNIT=SYSDA                     
//*                                                               
//SYSIN    DD *                                                   
  OPTION COPY                                                     
  INREC IFTHEN=(WHEN=GROUP,BEGIN=(1,1,CH,EQ,C'D'),               
                PUSH=(81:ID=4))                                   
  OUTFIL FNAMES=T1,BUILD=(1,84)                                   
  OUTFIL FNAMES=T2,REMOVECC,NODETAIL,BUILD=(9X),                 
                SECTIONS=(81,4,                                   
                TRAILER3=(81,4,COUNT-1=(M10,LENGTH=5)))           
//STEP0200 EXEC PGM=SORT                             
//SYSOUT   DD SYSOUT=*                               
//SORTJNF1 DD DSN=&T1,DISP=SHR                       
//SORTJNF2 DD DSN=&T2,DISP=SHR                       
//SORTOUT  DD SYSOUT=*                               
//SYSIN    DD *                                     
 JOINKEYS FILE=F1,FIELDS=(81,4,A),SORTED,NOSEQCK     
 JOINKEYS FILE=F2,FIELDS=(01,4,A),SORTED,NOSEQCK     
 JOIN UNPAIRED,F1                                   
 REFORMAT FIELDS=(F1:1,80,F2:5,5)                   
 SORT FIELDS=COPY                                   
 INREC IFOUTLEN=80,                                 
       IFTHEN=(WHEN=(3,5,CH,EQ,C'COUNT'),           
       OVERLAY=(3:81,5),HIT=NEXT),                   
       IFTHEN=(WHEN=ANY,                             
       FINDREP=(IN=C' ',OUT=C'',STARTPOS=3,ENDPOS=6))

SORTOUT had
Code:
D|4|111111111111|123.23|XYZ COMMERCE|NEW YORK CITY   
R|11111110444A|15022016|0.34|XYZ AFFILIATE|BOSTON   
R|11111110444A|15032016|2.44|XYZ AFFILIATE|BOSTON   
R|11111110444B|15022016|12.76|XYZ BROTHERS|LA CITY   
R|11111110444B|15032016|112.76|XYZ BROTHERS|LA CITY 
D|2|222222211111|73.23|ABC CONSTRUCTION|NEW YORK CITY
R|22222220111A|15022016|0.34|ABC CONSTRUCTION|BOSTON
R|22222220111B|15032016|71.63|ABC CONS.|BOSTON       
Back to top
View user's profile Send private message
sergeyken

New User


Joined: 29 Apr 2008
Posts: 37
Location: Maryland

PostPosted: Fri Sep 23, 2016 11:56 pm    Post subject:
Reply with quote

Note #1

Under no circumstances you can get the count of following records before you've read all those records.
You cannot guess in advance what is following next in your file?

It's the same as getting your salary in advance - before starting your job.

At least two steps are required to manipulate with record re-ordering.
But correct way would be creating the totals under your detail records.
Back to top
View user's profile Send private message
Atul Banke

New User


Joined: 08 Aug 2011
Posts: 5
Location: India

PostPosted: Mon Sep 26, 2016 5:10 pm    Post subject:
Reply with quote

Hi Arun,

Thanks a lot, it works perfectly fine as you suggested. I was also thinking that using WHEN=GROUP and then join should work but since I didn't know the SECTIONS parameter therefore could not get that working.

Thanks again!
Back to top
View user's profile Send private message
Bill Woodger

DFSORT Moderator


Joined: 09 Mar 2011
Posts: 7234

PostPosted: Tue Sep 27, 2016 1:07 am    Post subject: Reply to: Get Record count in summary record for each group of records
Reply with quote

At a brief view, you don't need two steps, all the processing of the first step can be specified in the JNF1CNTL and JNF2CNTL files for your single input file (used twice).
Back to top
View user's profile Send private message
Arun Raj

Moderator


Joined: 17 Oct 2006
Posts: 2163
Location: @my desk

PostPosted: Tue Sep 27, 2016 5:55 pm    Post subject:
Reply with quote

Atul Banke wrote:
Hi Arun,

Thanks a lot, it works perfectly fine as you suggested. Thanks again!
You're welcome!. Thanks for letting us know.
Back to top
View user's profile Send private message
magesh23586

Active User


Joined: 06 Jul 2009
Posts: 159
Location: Chennai

PostPosted: Tue Sep 27, 2016 9:00 pm    Post subject: Reply to: Get Record count in summary record for each group of records
Reply with quote

Arun,

As per Bill advice, We can do this in one step like below.

Code:

//STEP01  EXEC PGM=SORT                                             
//SORTJNF1  DD DSN=INPUT.FILE1,DISP=SHR                             
//SORTJNF2  DD DSN=INPUT.FILE2,DISP=SHR                             
//SORTOUT DD SYSOUT=*                                               
//SYSOUT  DD SYSOUT=*                                               
//SYSIN DD *                                                         
   OPTION COPY                                                       
   JOINKEYS FILE=F1,FIELDS=(81,6,A)                                 
   JOINKEYS FILE=F2,FIELDS=(81,6,A),SORTED,NOSEQCK                   
   
   JOIN UNPAIRED,F2 
                                               
   REFORMAT FIELDS=(F2:1,80,F1:88,5)   
                             
   INREC  IFOUTLEN=80,                                               
          IFTHEN=(WHEN=(3,5,CH,EQ,C'COUNT'),                         
          OVERLAY=(3:81,5,ZD,SUB,+1,EDIT=(IIIIT),81:5X),HIT=NEXT),
 
          IFTHEN=(WHEN=ANY,                                         
          FINDREP=(IN=C' ',OUT=C'',STARTPOS=3,ENDPOS=6)) 
           
//JNF1CNTL DD *                                                     
  INREC IFTHEN=(WHEN=INIT,                                           
                OVERLAY=(88:C'00001')),                             
                                                                     
        IFTHEN=(WHEN=GROUP,   
                BEGIN=(1,1,CH,EQ,C'D'),       
                PUSH=(81:ID=6))               
                SUM FIELDS=(88,5,ZD)           
//JNF2CNTL DD *                               
  INREC IFTHEN=(WHEN=GROUP,                   
                BEGIN=(1,1,CH,EQ,C'D'),       
                PUSH=(81:ID=6))               
                                     
Back to top
View user's profile Send private message
Arun Raj

Moderator


Joined: 17 Oct 2006
Posts: 2163
Location: @my desk

PostPosted: Tue Sep 27, 2016 9:49 pm    Post subject:
Reply with quote

magesh23586,

AFAIK it would require the same passes of data.

Besides you have used the SUM operator that requires 'sorting' of data. What I posted earlier was a SORT COPY, which does not require any sorting. Always the direction should be to avoid sorting of data whenever there is no need to. Saving resources should be the priority, we may not always gain by doing something in 'one step'.
Back to top
View user's profile Send private message
magesh23586

Active User


Joined: 06 Jul 2009
Posts: 159
Location: Chennai

PostPosted: Wed Sep 28, 2016 2:21 am    Post subject:
Reply with quote

Arun,

Here is the one step solution without sort icon_smile.gif. As I mentioned earlier we need not requires two steps

Code:

//SYSIN DD *                                                       
   OPTION COPY                                                     
   JOINKEYS FILE=F1,FIELDS=(81,6,A,103,1,A),SORTED,NOSEQCK         
   JOINKEYS FILE=F2,FIELDS=(81,6,A,103,1,A),SORTED,NOSEQCK         
                                                                   
   JOIN UNPAIRED,F1,F2                                             
                                                                   
   REFORMAT FIELDS=(F1:1,103,?,F2:1,103)                           
                                                                   
   INCLUDE COND=(207,1,CH,NE,C'C')                                 
                                                                   
   INREC  IFOUTLEN=103,                                           
          IFTHEN=(WHEN=(104,1,CH,EQ,C'2'),                         
          OVERLAY=(1:105,103)) 
                                   
   OUTREC IFTHEN=(WHEN=INIT,                                       
          BUILD=(97,7,1,80)),
                                     
          IFTHEN=(WHEN=INIT,                                       
          FINDREP=(IN=C' ',OUT=C'',STARTPOS=10,ENDPOS=13))         
   
   OUTFIL SECTIONS=(1,7,                                           
          TRAILER3=(8,80)),NODETAIL,REMOVECC                       

//JNF1CNTL DD *                                               
  INREC IFTHEN=(WHEN=INIT,                                     
                OVERLAY=(97:C'000001',C'A')),                 

        IFTHEN=(WHEN=GROUP,                                   
                BEGIN=(1,1,CH,EQ,C'D'),                       
                PUSH=(1:1,80,ID=6,SEQ=5)),
                   
        IFTHEN=(WHEN=(3,5,CH,EQ,C'COUNT'),                     
            OVERLAY=(3:87,5,ZD,SUB,+1,EDIT=(IIIIT)))           

//JNF2CNTL DD *                                               
  INREC IFTHEN=(WHEN=GROUP,
                BEGIN=(1,1,CH,EQ,C'D'),             
                PUSH=(81:ID=6,
                      97:SEQ=6)),
           
        IFTHEN=(WHEN=(1,1,CH,EQ,C'D'),
                OVERLAY=(103:C'C')),     

        IFTHEN=(WHEN=NONE,
                OVERLAY=(103:C'B'))
Back to top
View user's profile Send private message
Arun Raj

Moderator


Joined: 17 Oct 2006
Posts: 2163
Location: @my desk

PostPosted: Wed Sep 28, 2016 5:29 pm    Post subject:
Reply with quote

magesh23586,

There is no justification to build a 207 byte record in the REFORMAT to save a step. Keeping the Intermediate file lengths to the possible minimum is as important as getting rid of an unnecessary sort.

Learn to keep things simple and short. Like I mentioned earlier - everything in "one step" might not help always, and that should not be the prime objective of any application.
Back to top
View user's profile Send private message
enrico-sorichetti

Global Moderator


Joined: 14 Mar 2007
Posts: 10202
Location: italy

PostPosted: Wed Sep 28, 2016 6:18 pm    Post subject: Reply to: Get Record count in summary record for each group of records
Reply with quote

having a single step will make JOB management easier
in case of a failure RESTART from the failing step

for the example posted
a failure in step200 will imply a restart from step100

I would go for easier handling rather than for peephole optimisation
Back to top
View user's profile Send private message
Arun Raj

Moderator


Joined: 17 Oct 2006
Posts: 2163
Location: @my desk

PostPosted: Wed Sep 28, 2016 7:37 pm    Post subject:
Reply with quote

Below is a "one step" ICETOOL version of the DFSORT solution posted earlier.
Code:
//STEP0100 EXEC PGM=ICETOOL                   
//TOOLMSG  DD SYSOUT=*                         
//DFSMSG   DD SYSOUT=*                         
//IN       DD DISP=SHR,DSN ===>  Input Data set (FB/LRECL=80) 
//CTL2JNF1 DD DSN=&T1,DISP=(,PASS),UNIT=SYSDA 
//CTL2JNF2 DD DSN=&T2,DISP=(,PASS),UNIT=SYSDA 
//OUT      DD SYSOUT=*                         
//*                                           
//TOOLIN   DD *                               
COPY FROM(IN)       TO(CTL2JNF1)  USING(CTL1) 
SORT FROM(CTL2JNF1) TO(OUT)       USING(CTL2)
//*                                                   
//CTL1CNTL DD *                                       
  OPTION COPY                                         
  INREC IFTHEN=(WHEN=GROUP,BEGIN=(1,1,CH,EQ,C'D'),     
                PUSH=(81:ID=4))                       
  OUTFIL FNAMES=CTL2JNF1,BUILD=(1,84)                 
  OUTFIL FNAMES=CTL2JNF2,REMOVECC,NODETAIL,BUILD=(9X),
                SECTIONS=(81,4,                       
                TRAILER3=(81,4,COUNT-1=(M10,LENGTH=5)))
//CTL2CNTL DD *                                     
 JOINKEYS FILE=F1,FIELDS=(81,4,A),SORTED,NOSEQCK     
 JOINKEYS FILE=F2,FIELDS=(01,4,A),SORTED,NOSEQCK     
 JOIN UNPAIRED,F1                                   
 REFORMAT FIELDS=(F1:1,80,F2:5,5)                   
 SORT FIELDS=COPY                                   
 INREC IFOUTLEN=80,                                 
       IFTHEN=(WHEN=(3,5,CH,EQ,C'COUNT'),           
       OVERLAY=(3:81,5),HIT=NEXT),                   
       IFTHEN=(WHEN=ANY,                             
       FINDREP=(IN=C' ',OUT=C'',STARTPOS=3,ENDPOS=6))
Back to top
View user's profile Send private message
Bill Woodger

DFSORT Moderator


Joined: 09 Mar 2011
Posts: 7234

PostPosted: Thu Sep 29, 2016 3:32 am    Post subject: Reply to: Get Record count in summary record for each group of records
Reply with quote

magesh23586 contacted me yesterday after they noticed outside the 10-minute edit period that they'd included data which was not needed.

I'll try to get some edits done, maybe in the morning, as it is already late here.
Back to top
View user's profile Send private message
magesh23586

Active User


Joined: 06 Jul 2009
Posts: 159
Location: Chennai

PostPosted: Fri Sep 30, 2016 5:03 am    Post subject: Reply to: Get Record count in summary record for each group of records
Reply with quote

Arun Raj,

Arun Raj wrote:

Below is a "one step" ICETOOL version of the DFSORT solution posted earlier.


Though you do it in ICETOOL, your solution is 4 Pass solution (1 for sort copy + 3 for Joinkeys) , which is not at all recommended when there is 1 step( Joinkey 3 Pass) solution is possible.

what will happen when second JOINKEYS Fails for some reason, we need to start sort copy + joinkey again, which means we are wasting resource unnecessarily.

Bill, Please correct me, if i am wrong.


The following code would be optimal + simple, where I am just sorting 12 bytes of records which will be much better than a reading a complete file with Sections. Below is the post I requested Bill to update.


Code:

//SYSIN DD *                                                       
   OPTION COPY                                                     
   JOINKEYS FILE=F1,FIELDS=(1,6,A)                                 
   JOINKEYS FILE=F2,FIELDS=(81,6,A),SORTED,NOSEQCK                 
   JOIN UNPAIRED,F2                                                 
   REFORMAT FIELDS=(F2:1,80,F1:7,5)                                 
   INREC IFOUTLEN=80,                                               
          IFTHEN=(WHEN=(3,5,CH,EQ,C'COUNT'),                       
          OVERLAY=(3:81,5,ZD,SUB,+1,EDIT=(IIIIT),81:5X),HIT=NEXT),
          IFTHEN=(WHEN=ANY,                                         
          FINDREP=(IN=C' ',OUT=C'',STARTPOS=3,ENDPOS=6))           
//JNF1CNTL DD *                                                     
  INREC IFOUTLEN=12,                                               
        IFTHEN=(WHEN=INIT,                                         
                OVERLAY=(7:C'00001')),
                             
        IFTHEN=(WHEN=GROUP,                                         
                BEGIN=(1,1,CH,EQ,C'D'),                             
                PUSH=(1:ID=6))                                     
                SUM FIELDS=(7,5,ZD)     
//JNF2CNTL DD *                                     
  INREC IFTHEN=(WHEN=GROUP,
                BEGIN=(1,1,CH,EQ,C'D'),   
                PUSH=(81:ID=6))



Avoid Sorting, following code is still better than two step solution. where the combined record read is 116 bytes now.

Code:

//SYSIN DD *                                                       
 OPTION COPY                                                   
 JOINKEYS FILE=F1,FIELDS=(6,7,A),SORTED,NOSEQCK               
 JOINKEYS FILE=F2,FIELDS=(1,7,A),SORTED,NOSEQCK               
 JOIN UNPAIRED,F2                                             
 REFORMAT FIELDS=(F2:1,98,?,F1:1,17)                           
                                                               
 INREC  IFTHEN=(WHEN=(99,1,CH,EQ,C'B'),                       
        OVERLAY=(21:100,5)),                                   
        IFOUTLEN=98                                           
                                                               
 OUTREC IFTHEN=(WHEN=INIT,                                     
        FINDREP=(IN=C' ',OUT=C'',STARTPOS=21,ENDPOS=24))       
                                                               
 OUTFIL SECTIONS=(13,6,                                       
              TRAILER3=(19,80)),NODETAIL,REMOVECC,BUILD=(1,80)
//JNF1CNTL DD *                                             
  INREC IFOUTLEN=18,                                       
        IFTHEN=(WHEN=INIT,                                 
                OVERLAY=(12:C'A000001')),                   
                                                           
        IFTHEN=(WHEN=GROUP,                                 
                BEGIN=(1,1,CH,EQ,C'D'),                     
                PUSH=(1:SEQ=5,ID=6)),                       
                                                           
        IFTHEN=(WHEN=INIT,                                 
                OVERLAY=(1:1,5,ZD,SUB,+1,EDIT=(IIIIT)))     
//JNF2CNTL DD *                                             
  OPTION COPY                                               
  INREC IFTHEN=(WHEN=INIT,                                 
                BUILD=(6X,C'B',11X,1,80)),                 
                                                           
        IFTHEN=(WHEN=GROUP,                                 
                BEGIN=(19,1,CH,EQ,C'D'),                   
                PUSH=(1:ID=6,13:SEQ=6)),                   
                                                           
        IFTHEN=(WHEN=(19,1,CH,EQ,C'D'),                     
               OVERLAY=(7:C'A'))   
Back to top
View user's profile Send private message
Arun Raj

Moderator


Joined: 17 Oct 2006
Posts: 2163
Location: @my desk

PostPosted: Fri Sep 30, 2016 6:19 pm    Post subject:
Reply with quote

Bill Woodger wrote:
So any solution involving SORT is absurd: wasteful of resources. Unless absolutely necessary :-)
magesh23586,

I would assume that the above note by Bill in one of the old topics still holds true for your "optimum + simple" solution-1, which involves SORT and SUM.

Moreover, I would think we don't have to restart/abend on a daily basis. So if the copy solution runs fine for 99% of the time, I don't see any problem with that, unless the job fails every day.

Edit: Your second "solution" still has unused bytes if you have not realized it yet.
Back to top
View user's profile Send private message
Bill Woodger

DFSORT Moderator


Joined: 09 Mar 2011
Posts: 7234

PostPosted: Fri Sep 30, 2016 9:39 pm    Post subject: Reply to: Get Record count in summary record for each group of records
Reply with quote

Problem is, the context of the original from which the quote is taken is the ability to use the OUTFIL reporting features. "Don't use SORT just to use SUM". In JNFnCNTL, there is no possibility to use the OUTFIL reporting features. But in itself, perhaps not enough to use SUM.

It's not the SUM itself that is the problem, it is doing a SORT just to be able to do a SUM, because SORTing is an extensive operation.

I think from Arun's original solution, drop the first OUTFIL and do the appending in the JNF1CNTL. Avoids the "write", and the BUILD (which I can't see why it was needed).

For sure I've seen and used the SUM in JNFnCNTL for this type of task. A separate step I think on balance is going to be a more efficient way to do it than by using the SUM.

From both the suggested solutions, I have another idea. Have to check it first...
Back to top
View user's profile Send private message
Arun Raj

Moderator


Joined: 17 Oct 2006
Posts: 2163
Location: @my desk

PostPosted: Fri Sep 30, 2016 11:19 pm    Post subject: Re: Reply to: Get Record count in summary record for each group of records
Reply with quote

Regarding the solution-2 by magesh23586, we don't really need to complicate things by adding another identifier 'A'.

I would do something like this, if this has to be a "one step" DFSORT thing.
Code:
//STEP0100 EXEC PGM=SORT                                             
//SYSOUT   DD SYSOUT=*                                               
//SORTJNF1 DD DSN= >> Input data set      (FB/80)                   
//SORTJNF2 DD DSN= >> Same Input data set (FB/80)                   
//SORTOUT  DD DSN= >> Output data set     (FB/80)                   
//*                                                                 
//SYSIN    DD *                                                     
  JOINKEYS FILE=F1,FIELDS=(81,4,A,01,01,A),SORTED,NOSEQCK           
  JOINKEYS FILE=F2,FIELDS=(01,4,A,10,01,A),SORTED,NOSEQCK           
  JOIN UNPAIRED,F1                                                   
  REFORMAT FIELDS=(F1:85,5,1,80,F2:5,5)                             
  SORT FIELDS=COPY                                                   
  INREC IFTHEN=(WHEN=(8,5,CH,EQ,C'COUNT'),                           
        OVERLAY=(8:86,5),HIT=NEXT),                                 
        IFTHEN=(WHEN=ANY,                                           
        FINDREP=(IN=C' ',OUT=C'',STARTPOS=08,ENDPOS=11))             
  OUTFIL REMOVECC,NODETAIL,SECTIONS=(1,5,TRAILER3=(6,80)),BUILD=(80X)
/*                                                                   
//JNF1CNTL DD *                                                       
  INREC IFTHEN=(WHEN=GROUP,BEGIN=(1,1,CH,EQ,C'D'),                   
                PUSH=(81:ID=4,SEQ=5))                                 
/*                                                                   
//JNF2CNTL DD *                                                       
  INREC IFOUTLEN=10,                                                 
        IFTHEN=(WHEN=GROUP,BEGIN=(1,1,CH,EQ,C'D'),                   
                PUSH=(01:ID=4,SEQ=5)),                               
        IFTHEN=(WHEN=INIT,OVERLAY=(5:5,5,ZD,SUB,+1,M10,LENGTH=5,C'D'))
/*   

My personal preference is not to generate duplicate records and then having to roll up later. I would have to run a performance comparison between the 2 step DFSORT and the above one. I will post it here when I have it.
Back to top
View user's profile Send private message
Rohit Umarjikar

Senior Member


Joined: 21 Sep 2010
Posts: 1610
Location: NY,USA

PostPosted: Sat Oct 01, 2016 1:47 am    Post subject:
Reply with quote

Just a thought, between these big performance talks and may get avoided with simple approach.
1.Assuming the Header (D) is unique always.
Here is my approach.
a. I am not sure if sequence number is required but in case to sort back or its original order I have added, also '1' for reach record
b.Whenever we get a hit for Header , push the entire header for reach record (append).
c.group them (pushed headers) and count-This gives the details record count to overlay
d.Overlay the count in step c ,when its ONLY header(D) and include only records which has a space at 51:1
I stopped right here and got to go, may be someone can modify to finsh.
Code:
//SORTIN  DD *     
D123           
R343           
R343           
D454           
R343           
//SYSIN DD *
INREC  IFOUTLEN=80,                                 
IFTHEN=(WHEN=INIT,OVERLAY=(30:C'1',40:SEQNUM,8,ZD)),
IFTHEN=(WHEN=GROUP,                                 
        BEGIN=(1,1,CH,EQ,C'D'),PUSH=(20:1,5))       
OUTFIL REMOVECC,FNAMES=T1,                         
         SECTIONS=(20,5,                           
         TRAILER3=(50:COUNT-1=(M10,LENGTH=1)))     
Back to top
View user's profile Send private message
Arun Raj

Moderator


Joined: 17 Oct 2006
Posts: 2163
Location: @my desk

PostPosted: Sat Oct 01, 2016 2:10 am    Post subject:
Reply with quote

Hi Rohit,

Just in case if you missed to notice, the approach described above to arrive at the detail record count is already used in one of the tested working solutions posted so far in this topic.
Rohit Umarjikar wrote:
c.group them (pushed headers) and count-This gives the details record count to overlay
d.Overlay the count in step c ,when its ONLY header(D) and include only records which has a space at 51:1
I could not comprehend how this piece can possibly get to the OPs expected result.
Back to top
View user's profile Send private message
View previous topic :: :: View next topic  
Post new topic   Reply to topic    IBMMAINFRAMES.com Support Forums -> DFSORT/ICETOOL All times are GMT + 6 Hours
Goto page 1, 2  Next
Page 1 of 2

 

Search our Forum:

Similar Topics
Topic Author Forum Replies Posted
No new posts Removing Duplicates based on certain ... chandracdac DFSORT/ICETOOL 1 Fri Dec 09, 2016 4:40 am
No new posts Limit duplicate records in the SORT pshongal SYNCSORT 6 Mon Nov 21, 2016 12:54 pm
No new posts How to update a portion of text in a ... Bill Woodger DFSORT/ICETOOL 25 Wed Nov 09, 2016 9:41 pm
No new posts How to split the records using the am... vnktrrd DFSORT/ICETOOL 24 Fri Oct 28, 2016 7:33 pm
No new posts Sort records based on numeric field. Alks SYNCSORT 2 Wed Oct 19, 2016 10:14 pm


Facebook
Back to Top
 
Mainframe Wiki | Forum Rules | Bookmarks | Subscriptions | FAQ | Tutorials | Contact Us