PDI 3.2.0 went stable

Dear Kettle friends,

Because there were so many people wondering what GA or even “Generally Available” meant, we renamed the thing to “Stable”.  So the stable production ready version of Pentaho data integration 3.2.0 now is called 3.2.0-stable.  Go figure.

In any case, Pentaho Data Integration 3.2.0 is generally available on sourceforge over here.

We changed a lot of things in that release, from small visual improvements in the hop color schema to larger changes behind the covers with things like named parameters.

Other pet peeves that changed are for example the Table Output step where we now (finally!) are capable of specifying individual fields to be stored.  The Dimension Lookup/Update step got an upgrade as well with support for cache pre-loading, alternative algorithms for date ranges, current row flag and much more.  Obviously, we got a whole batch of new steps.  Simple steps were added like “Replace values in string” or the ultra-fast User Defined Java Expression.

For an overview of the changes, have a look at the What’s new in 3.2 page on the wiki.

Enjoy this release & thank you all for your continued support and help,


12 thoughts on “PDI 3.2.0 went stable”

  1. Matt,
    I have been wondering, since you have taken up a gui-designer approach, why havent opted to provide a scaled down version of a web designer. It is understood that monitoring live processes is not easy. But running the kettle as a web application is something that would be a common practice for most of the vendors who have already built their j2ee stack.


  2. Jana, besides the fact that nobody else is doing it in thin client, ETL is not something that needs to be done by a lot of people. Besides, you can WebStart Kettle on your J2EE stack just fine.
    All that being said, slowly but surely we are pushing in a thin client direction.


  3. Matt,
    I see that more as an oppurtnuity for better alignment(that nobody else has it). As you mentioned Webstarting kettle is definitely an option, the ability to perform source code management for kettle definitions has to be still addressed(I think I am just expressing my discomfort on javawebstart for enterprise tools like kettle, if it could have been some player like thing I am definitely ok with it). Kettle does relieve developers by letting save definitions in its own repository, I am game with it if we could make comply with j2ee standards like jcr. I think putting kettle on the web gives us the oppurtunity to align our efforts in one direction. On a final note, it looks like you have already started in that direction. Will be waiting to hear on that.

    BTW, I have been associated to kettle from 2.5.0 and would like to take the oppurtunity to say, you have done very good job and saved my b*#$% a many number of times.

  4. Jana, the discussion is held elsewhere, but it looks like the next release will be pushing towards the CMIS standard for document/file management and versioning.
    In another area, the change is more transparent, but since version 3.0 we’re using XUL for the Spoon framework. We’re also going to convert a few steps over, keep pushing in that direction. Pentaho already has XUL-Swing, XUL-SWT *and* XUL-GWT bridges making it easy for us to be framework independent in the future.
    Obviously it would be somewhat cool to have a thin client interface to transformations and jobs. However, the complete set of functionality, ease of use, etc can’t be duplicated yet in a thin client environment. I’m sure that in a few years, when the complete feature set WILL be available on a thin client, we’ll be ready to transition.

  5. Hi,
    why is in “3.2.0” XML Input component “deprecated” ? this component is good for expanding multilevel XML files. It´s only one exist way to read it…


  6. Hi Matt,

    I’m using the SQL Script job step… In the bug tracker it’s said that the issue with semicolon is fixed… it’s true if you put the sql script directly in the step but if you use the “sql from file” option it doesn’t work. Would it be fixed soon?


  7. If you want to get a bug fixed with priority, there is only one way: get a support contract.
    Please stop spamming this blog. Reserve these comments for the JIRA bug tracking system.


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.