In case of sequential files, it is not possible to directly read them from the last record until we read the records to a table and then go ahead to read from the last record.

in case of KSDS, we can move high-values to the variable and then read it from the highest value.

Hi,

 It is possible through jcl. For example if the input file contains 80 record length then

 In sort, INREC=(1:1,80,81:SEQNUM,6,ZD),SORT FIELDS=(81,6,ZD,D),OUTREC=(1:1,80)

 1) Using inrec we need to generate the sequential number starts from position 81.

     For example

           1) First 1-80 bytes contains the actual record and from position 81 we need to generate sequential 

 number using inrec i.e 000100, 000200 upto 50 records.  

           2) After generating all those numbers, sort we takes place in descending order.

           3) Using outrec we have to move the first 80 bytes to the output file.

 After doing this sort in first step in jcl, you can read the output file in second step

 Thanks

 Karthikkumar.P