Hitachi Vantara Pentaho Community Forums
Results 1 to 3 of 3

Thread: Insert / Update: Unable to compare with value...

  1. #1
    Join Date
    Feb 2013

    Default Insert / Update: Unable to compare with value...

    I am using Kettle version 5.0.1.-stable.

    An Insert / Update step sporadically fails with this error message:

    BigNumber : Unable to compare with value [BigNumber(17, 2)]

    In the beginning it was failing very sporadically. The Transformation is re-started in a job after 10 seconds again.
    After a while the row is written to the target.
    In the meantime it's happening with every 10th record.

    We are experiencing the problem in production only (Ubuntu 12.04.4 LTS) .
    It does not happen on my Windows development machine nor in our Ubuntu Beta environment.
    Even if I copy the same source row into the beta database it will not fail.

    Because of the fact that the record IS written after several retries (sometimes it takes hours, but finally the row will be written), I assume, that the error is not in the Update / Insert statement itself.
    Can it be something in the synchronization/preparation in the dataflow? That the data type isn't clear in the moment when the update is checked?
    I put a Block until step in front, to wait for all lookups and calculations, but it didnt change anything.

    The Job did run perfectly for some weeks.
    Then we added some calculated currency fields to the target table (decimal 17,2).
    Each row always had (and has) an individual (local) currency.
    In a first step I convert them to USD after a Lookup step for the conversion.
    In a second step I get EUR, GBP, INR and JPY to USD courses in 1 Table Input, to calculate these values.
    Since we a doing this, we have the Update problems.

    What can cause that it sometimes works and sometimes does not for the same imput rows?

    Thanks for any help.
    Last edited by Nidderauer; 03-20-2014 at 10:17 AM.

  2. #2
    Join Date
    Nov 1999


    What's the complete error? Did you get any stack trace?

  3. #3
    Join Date
    Feb 2013


    Oh, that's really embarrassing for me

    I changed the transformation about 40 times during the last 2 weeks, without any success.
    But the day, when I opened this thread, by mistake I only replaced the ktr file in our beta environment and not on the production server.
    So it still seemed to be wrong, but finally I fixed the error...

    I still don't understand why the errors happened sporadically and not consistent.
    But what I can tell are the last few changes I did, before I noticed that I was checking the wrong system and that the problem was fixed:

    I have 2 paths (maybe path is the wrong word, by I just don't know a better one) through the transformation, one for zero prices and one for the regular ones. In case of zero I do not need the lookup stuff and I just set all needed values to zero.
    Unfortunately I did not notice, that the order of the variables was not exactly the same in the 2 paths.
    The wrong order was indicated by the Verify tool.
    I changed the order and additionally some precision info of some variables until the output fields before the 2 paths met each other were absolutely identical.

    Is it possible that the transformation could get corrupt if the order of the fields gets mixed up in 2 separated and later joining paths?

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
Privacy Policy | Legal Notices | Safe Harbor Privacy Policy

Copyright © 2005 - 2019 Hitachi Vantara Corporation. All Rights Reserved.