SPSS Statistics

 View Only
  • 1.  reading files before they are actually written

    Posted Fri August 20, 2021 01:22 PM

    Hello,

     

    I have somewhat of a curious problem.

    I have a set of SPSS macros that write and then read many data files. When the target directory is a cloud, occasionally the program will tell me a file does not exist, even though 1 or 2 lines earlier I have confirmation that the file was indeed written successfully. All this happens within SPSS. When this happens, and I rerun the program with no changes whatsoever, the program usually runs with no problem. To avoid this I point to a directory on my local machine, and do not run into the problem at all. But I have users of the macros that still use a cloud as their target and run into this problem of looking for a file that should be there, but that it is not.

    More recently, in a newer macro that uses the HOST command to start an R session that creates a csv file, when SPSS reads the csv file 2 or 3 commands after HOST was done, it picked up 206 cases, but in looking at the csv file I see 5548 cases that should have been read by SPSS. There was no error or warning message reported in R or SPSS. I suspect this issue is related to the one described earlier.

     

    Is there a way to prevent "this" from happening? I suspect "this" is the program/system sending a file to be written but remaining in some cached space waiting for the proper time to be saved, while subsequent commands are executed and failing to find the file that has yet to be written.

     

    Thanks!

     



    This e-mail and any files transmitted with it may contain privileged or confidential information. It is solely for use by the individual for whom it is intended, even if addressed incorrectly. If you received this e-mail in error, please notify the sender; do not disclose, copy, distribute, or take any action in reliance on the contents of this information; and delete it from your system. Any other use of this e-mail is prohibited.


    Thank you for your compliance.



    #SPSSStatistics


  • 2.  RE: reading files before they are actually written

    IBM Champion
    Posted Fri August 20, 2021 02:03 PM
    So it seems that the file directory is not immediately updated.  Maybe you could put in a few seconds of sleeping at strategic points.  I wrote a trivial extension command, STATS SLEEP DELAY=time-in-seconds.

    It's not fully packaged and is not on the Extension Hub, but I can post the files on my OneDrive site if you want to try it.

    --





  • 3.  RE: reading files before they are actually written

    Posted Fri August 20, 2021 02:09 PM

    Sure, I can give that a try and see how it works.

    Thanks!

     






  • 4.  RE: reading files before they are actually written

    IBM Champion
    Posted Fri August 20, 2021 03:15 PM
    Download this file and unzip to a location where Statistics can find it.  Run SHOW EXT to see where it might look

    I would start with, say, 5 seconds and maybe cut it down after that to see what works, assuming that latency is the problem.
    STATS SLEEP DELAY=5.

    --