Hitachi Vantara Pentaho Community Forums
Results 1 to 5 of 5

Thread: None of the data-factories was able to handle the query...

  1. #1
    Join Date
    Feb 2009
    Posts
    14

    Question None of the data-factories was able to handle the query...

    This is an odd problem.

    PDI - [4.4.0-stable]
    PRD - [3.9.1-GA]



    I have been running a job for many months successfully using PDI that would use a Pentaho Reporting Output transformation to generate close to 200 reports at a time. I had a request from my user base to change the sort order of some of the items in my report. I accomplished this by going into the Pentaho Report Designer and modifying the query to order by an additional field. No problems. I tested the report with print preview in PRD, everything looked good...

    When my PDI job ran later that night, my users found that all the reports were empty...

    After I checked the logs, I am finding these errors...

    Code:
    WARN  10-05 16:21:37,082 - Deprecated behavior: None of the data-factories was able to handle the query 'Detail'. Returning empty tablemodel instead of failing hard.
    WARN  10-05 16:21:37,082 - Be aware that the default for this setting will change in version 0.8.11. To avoid this warning, make sure that all data-sources are properly configured and that no report references illegal queries.
    WARN  10-05 16:21:37,091 - Deprecated behavior: None of the data-factories was able to handle the query 'Summary'. Returning empty tablemodel instead of failing hard.
    WARN  10-05 16:21:37,091 - Be aware that the default for this setting will change in version 0.8.11. To avoid this warning, make sure that all data-sources are properly configured and that no report references illegal queries.
    A quick search in these forums result in a couple of different possibilities:

    This particular post indicates that the query may not be tied to the report properly.
    http://forums.pentaho.com/showthread...dle-this-query

    Ive checked the report (and 1 subreport I have) and they both are set ok. Moreover, the report works fine when I run it in PRD.

    This particular post talks about incorrect versions of PDI vs PRD.
    http://forums.pentaho.com/showthread...le+this+query.

    I confirmed that 3.9.1-GA should work correctly with 4.4.0-stable.

    I've went as far as rebuilding the report from scratch. No dice.


    Thoughts?

  2. #2
    Join Date
    Mar 2003
    Posts
    8,085

    Default

    I bet PDI embeds an old version of the reporting engine. Check PDIs lib directory for pentaho-reporting-engine-classic jars and their version number.

    Up until recently PDI had a different release cycle and a manual dependency management and thus library updates were not necessarily in sync with the latest BI-Server or PRD release of that time.
    Get the latest news and tips and tricks for Pentaho Reporting at the Pentaho Reporting Blog.

  3. #3
    Join Date
    Feb 2009
    Posts
    14

    Default

    Quote Originally Posted by Taqua View Post
    I bet PDI embeds an old version of the reporting engine. Check PDIs lib directory for pentaho-reporting-engine-classic jars and their version number.

    Up until recently PDI had a different release cycle and a manual dependency management and thus library updates were not necessarily in sync with the latest BI-Server or PRD release of that time.
    I was looking in the "\libext\reporting\" directory and found what I believe you are referring to. Would you expect that the Jar files would be a different (older) versions? It looks to me that they were all ok but I may be not understanding what you were referring to.
    Attached Images Attached Images  

  4. #4
    Join Date
    Mar 2003
    Posts
    8,085

    Default

    Hmm, looks totally sane. Can you ramp up the logging for "org.pentaho.reporting.engine.classic.core.modules.parser." to debug and run one report? This should give us some insight on whether the datasources are parsed wrongly.

    Also: Do you run the transformation in Spoon or is this on either a slave node or an bi-server (and called/scheduled there)? If it is there, then check their classpath as well, as each installation is different.

    There was only one change that affected data-sources in such a way that queries could be lost. In 3.9.0 we introduced query-scripting, which required a change in the XML format. The 3.9.x parser accepts both the old and the new format, while a 3.8.3 parser naturally on knows about the old format. The most common error where we have seen this issue is when people create reports in PRD-3.9 to let it run in PRD-3.8.3.

    This ultimately led me to (a) write a blog-post about time travel and (b) add a error check in PRD-3.9.1 rejecting newer PRPT versions in the future.

    If the wrong-version used error is there, you will see a message along the lines of

    Code:
    [WARN] AbstractXmlReadHandler - Unknown tag <http://jfreereport.sourceforge.net/namespaces/datasources/sql:some-tag-name>: Start to ignore this element and all of its childs.  [Location: Line=xx Column=yy]
    Get the latest news and tips and tricks for Pentaho Reporting at the Pentaho Reporting Blog.

  5. #5
    Join Date
    Feb 2009
    Posts
    14

    Default

    Quote Originally Posted by Taqua View Post
    Hmm, looks totally sane. Can you ramp up the logging for "org.pentaho.reporting.engine.classic.core.modules.parser." to debug and run one report? This should give us some insight on whether the datasources are parsed wrongly.

    Also: Do you run the transformation in Spoon or is this on either a slave node or an bi-server (and called/scheduled there)? If it is there, then check their classpath as well, as each installation is different.

    There was only one change that affected data-sources in such a way that queries could be lost. In 3.9.0 we introduced query-scripting, which required a change in the XML format. The 3.9.x parser accepts both the old and the new format, while a 3.8.3 parser naturally on knows about the old format. The most common error where we have seen this issue is when people create reports in PRD-3.9 to let it run in PRD-3.8.3.

    This ultimately led me to (a) write a blog-post about time travel and (b) add a error check in PRD-3.9.1 rejecting newer PRPT versions in the future.

    If the wrong-version used error is there, you will see a message along the lines of

    Code:
    [WARN] AbstractXmlReadHandler - Unknown tag <http://jfreereport.sourceforge.net/namespaces/datasources/sql:some-tag-name>: Start to ignore this element and all of its childs.  [Location: Line=xx Column=yy]

    Eureka!

    For some reason I had it stuck in my head that the server that my main PDI installation sits was where the jobs were running from. In fact, I have a scripting server that sits and runs jobs (many jobs in fact) locally which had the 3.8 version of PDI installed! Whew! Not sure why my brain couldn't make that connection but something in your last post "clicked" for me.

    Thank you so much for the advice.

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.