Joining two files with SORT.

[Roohi N]

Hi,

I have two files which have just one-one record each. I need to combine them in a specific way. I shall try to explain them with example:

Let's say file one is: my.input.file1 with data 12345
file second is: my.input.file2 with data abcdef09876543

I want my output file should be as: 12345abcdef09876543

Can I do this using some SORT card?

[Roohi N]

The files are both of logical length 80 and and fixed blocked.

I might get these files as variable block, can we still do it in SORT with variable block file?

[Magesh_j]

Roohi,

what is the output LRECL ?

Whether fields you have are in fixed length or length varies ?

Assuming output LRECL=80 and All the fields have fixed length then use the below code and you have only one record in each file.

File 1(1,5) and File2 (1,14)

Code: Select all

//SORT010    EXEC PGM=SORT                                  
//SORTIN  DD DSN=YOUR FIRST FILE                            
//        DD DSN=YOUR SECOND IFLE                           
//SORTOUT DD SYSOUT=*                                       
//SYSOUT DD SYSOUT=*                                        
//SYSIN DD *                                                
  OPTION COPY,STOPAFT=2                                     
  INREC IFTHEN=(WHEN=GROUP,RECORDS=2,PUSH=(81:1,5,SEQ=1))   
                                                            
  OUTFIL INCLUDE(86,1,CH,EQ,C'2'),BUILD=(81,5,1,14,80:X)    
/*                                                          

On you next question on Variable file, what is the maximum length of input record ? And what is the maximum length of output record ?

[nicc]

There are NO files. They are data sets.

[Magesh_j]

Nic,

Mistake, agreed Datasets.

Thanks
Magesh

[Roohi N]

what is the output LRECL ?Whether fields you have are in fixed length or length varies ?Assuming output LRECL=80 and All the fields have fixed length then use the below code and you have only one record in each file.File 1(1,5) and File2 (1,14)

The datasets are both of logical length 80 and and fixed blocked.

Length of field is fixed for my case but the dataset's can be of RECFM=VB (this is a condiction we are expecting but as of now they are FB. THat's why I'm thimking for VB soluction too).

For VB, adding (1,4) should work but I'm still to get a VB file. Please advise.

[nicc]

But what is the output dataset LRECL?

[zprogrammer]

Roohi,

I would suggest : Overlay sequence numbers in both files and then do a join keys

[Magesh_j]

Roohi,

I would suggest : Overlay sequence numbers in both files and then do a join keys

Joinkey consumes three pass. where the solution which i have provided consume only two pass.

For this situation join key is not at all recommended.

Roohi N.

For VB yes, you should use 1,4 (RDW), Also you need to know the maximum record length.

[zprogrammer]

For this situation join key is not at all recommended.

Could you please prove it

[Magesh_j]

zprogrammer,

Check the spool for Joinkey, you will see three outputs.

Code: Select all

SYSOUT  
JNF1JMSG
JNF2JMSG

JNF1 => 1 Pass.
JNF2 => 1 Pass.
After Join => 1 pass.

ideally any DFSORT Joinkey will have three pass irrespective of JNF1CNTL/JNF2CNTL statements.

Also test my solution and you will see only one SYSOUT( ideally one pass) but since it is two file we concatenate it is considered as two pass solution.

Regards,
Magesh

[zprogrammer]

Magesh_j,

You are yet to answer my question => why is Join keys not recommended for this situation ?

[Magesh_j]

zprogrammer,

Because three pass is unnecessary, where something can be achieved in two pass which would be an optimal solution.

thanks
Magesh

[enrico-sorichetti]

the TS stated clearly that

I have two files which have just one-one record each.