How to count No. of Records in Dataset using Sort?

[harish123]

Hi

Using JCL/SORT how can we find the total number of records in a dataset. Are there any tools specific for this or can we use SYNCSORT to do this. If so how to do that?

Can some one help me.

[Robert Sample]

There are multiple ways to do this. One way would be to add a sequence number field to the record. Or add a constant 1 to the record and sum them. Or just use a COPY and look at the job output to see how many records SORT found.

The answer to your question partly depends upon what you meant -- do you want the count in a trailer record, or just printed, or stored on the last record, or something else?

[BobThomas]

Should not REXX be used here?

[Robert Sample]

If the data set has about 2 billion records, I would NOT recommend using REXX.

[Anuj Dhawan]

I'm in agreement with Robert - for huge data sets, (ad Robert has already indicated what is "Huge") - REXX is not a good choice.

[harish123]

Thanks Robert and Anuj.

I intend to use this for my daily work - it's not a real requirement though. I'd like to use Sort for this.

[Anuj Dhawan]

As Robert has said, there are multiple ways of doing it. Keeping the latest post from you in mind - one of follwoing should help you:

Code: Select all

//STEP010  EXEC PGM=ICETOOL                                            
//TOOLMSG  DD  SYSOUT=*                                                  
//DFSMSG   DD  SYSOUT=*                                                  
//DD01     DD  DSN=TAPE_FILE_NAME,                                      
//             DISP=SHR                                                  
//TOOLIN   DD  *                                                        
   COUNT FROM(DD01)                                                      
/*                                                                      
//*  

Check DFSMSG for DD01, in S.ST.

You can alos use this code, this is rather straight and can be used directly in an Application where (merely) the reocrd count might be of interest. As the LENGTH=4 is specified in SYSIN, this can be used for input records=9999, you can amend it accordingly :

Code: Select all

//STEP010  EXEC PGM=SORT              
//SYSOUT   DD  SYSOUT=*               
//SORTIN   DD  DSN=your.file.name,
//             DISP=SHR               
//SYSIN    DD  *                      
  OPTION COPY                         
   OUTFIL NODETAIL,REMOVECC,          
   TRAILER1=(COUNT=(M11,LENGTH=4))    
/*                                    
//SORTOUT  DD  SYSOUT=*               
//*

Hope this helps.

[harish123]

Thank you Anuj. This is very useful. I don't need this for some production requirement but to make my daily work life easy.

Thanks again.

[Anuj Dhawan]

Glad you find it helpful...

[BobThomas]

If the data set has about 2 billion records, I would NOT recommend using REXX.

Thanks - that make sense.