Shell Programming and Scripting

I've searched previous forums but could not find the answer to my question.

What determines how the "Work space used (bytes)" line item in the SyncSort Statistics appears? For instance I have some that show:


Records read: 3,273,645 Data read (bytes): 1,132,681,170
Records sorted: 3,273,645 Data sorted (bytes): 1,132,681,170
Records output: 3,273,645 Data output (bytes): 1,132,681,170
Input record length: 346 Output record length: 346
Work space used (bytes): 0

And also this:

Records read: 17,233,653 Data read (bytes): 2,240,374,890
Records sorted: 17,233,653 Data sorted (bytes): 2,240,374,890
Records output: 17,233,653 Data output (bytes): 2,240,374,890
Input record length: 130 Output record length: 130
Work space used (bytes): 2,278,948,864

Thank You, in advance.

Assuming this is for the same syncsort statement but different datasets, it is because the working memory available to the process was sufficient to complete the sort in memory. Based on the output, that may be a bad assumption.

Workspace is allocated on disk, is slower. Syncsort is optimized to choose the fastest sort method for a given dataset.

Thanks Jim, that makes sense the smaller file would use the working memory.

Performance is not an issue for us however a recent process failed due with "all available work space is exhausted". The workspace directory is 12 gig so we were suprised for the failure.

We are pulling back reports to see what the workspace directory usage was during that time.

Thanks again