Re: Any way to improve on CPYTOIMPF's speed?

[Dan <dan27649@xxxxxxxxx>]

Need to do record selection on the first copy (to the work file) since
CPYTOIMPF doesn't allow for record selection.  Still, that's only 13
minutes right there.  The CPYTOIMPF needs an hour and 39 minutes to convert
it to .CSV???  Good grief.

I compiled a Q&D program to start debug and issue the command again.  Not
seeing anything significant in the job log:
  Query options retrieved file QAQQINI in library QUSRSYS.
  Query options retrieved file QAQQINI in library QUSRSYS.
  DESCRIBE of prepared statement SQLSTATEMENT000003 completed.
  PREPARE of statement SQLSTATEMENT000003 completed.
  Ownership of object Q2T2TT02 in QTEMP type *FILE changed.
  Alias Q2T2TT02 created in QTEMP.
  Query options retrieved file QAQQINI in library QUSRSYS.
  The query access plan has been rebuilt.
  DESCRIBE of prepared statement SQLSTATEMENT000003 completed.
  PREPARE of statement SQLSTATEMENT000003 completed.
  Query options retrieved file QAQQINI in library QUSRSYS.
  **** Starting optimizer debug message for query .
  Arrival sequence access was used for file IQX002H.
  **** Ending debug message for query .
  ODP created.
  Blocking used for query.
  Cursor SQLCURSOR000000003 opened.
  1 rows fetched from cursor SQLCURSOR000000003.
  ODP deleted.
  Cursor SQLCURSOR000000003 was closed.
  Query options retrieved file QAQQINI in library QUSRSYS.
  The query access plan has been rebuilt.
  DESCRIBE of prepared statement SQLSTATEMENT000003 completed.
  PREPARE of statement SQLSTATEMENT000003 completed.
  Query options retrieved file QAQQINI in library QUSRSYS.
  **** Starting optimizer debug message for query .
  Arrival sequence access was used for file IQX002H.
  **** Ending debug message for query .
  ODP created.
  Blocking used for query.
  Cursor SQLCURSOR000000003 opened.
  273 rows fetched from cursor SQLCURSOR000000003.

^ That last line is repeated many times as the job continues to run.

Rob, did your thread recommend a faster alternative?

- Dan

On Thu, Jan 26, 2017 at 1:33 PM, Rob Berendt <rob@xxxxxxxxx> wrote:

> It's me!!!
> Want to see all the overhead of CPYTOIMPF?  Try this command before
> running it
> STRDBG PGM(*NONE) UPDPROD(*YES)
> Then look at your joblog at the SQL!
>
> That and you're doubling your I/O's by writing it to a work file and then
> writing it out to a stream file.