alias in pipeline with wrong content

    This site uses cookies. By continuing to browse this site, you are agreeing to our Cookie Policy.

    • alias in pipeline with wrong content

      Hi all,

      my config: ETL Server 1.3 Palo Server 3.0.. Linux Cent OS, using ETL Client to start jobs..

      I have a table customer with customer_id and customer_name..
      So: customer_id is the element name of my dimension customers and customer_name is alias for that element..

      In a pipeline I want to attributes..

      <pipeline name="Customer_Pipeline" type="Tree">
      <sources>
      <source nameref="Customer_Source" />
      </sources>
      <target>
      <levels>
      <level>
      <input nameref="CustomerID" />
      <attributes>
      <attribute name="CustomerName">
      <input nameref="CustomerName" />
      </attribute>
      </attributes>
      </level>
      </levels>
      </target>
      </pipeline>



      Using <attribute> tag in <attributes>.. nothing happens.
      Using <alias> tag in <attributes>.. the names are visible.

      Strange that <attribute> is not working.. doesn't make sense for me.. but I will forget that.

      But now my problem:
      Image today I have 5 more Customers than yesterday. So I start a job to add new customers before updating my cube. After that I can see the new customer_id s but all alias names are the same. They have all the latest alias name:
      Pos Element Name CustomerName
      ...
      15 id15 customername15
      16 id16 customername16 (latest entry before update)
      17 id17 customername16
      18 id18 customername16
      19 id19 customername16
      20 id20 customername16
      21 id21 customername16

      I tried different modes for my export (update, add).. all the same..

      Any idea why? Would help me a lot

      Thanks
      sivgin

      The post was edited 1 time, last by sivgin ().

    • Ok.
      Using <attribute> tag in <attributes> is working on excel add-in...
      but <attribute> didn't help me out with CustomerName...

      and of course i am simply using the exports ..

      <export name="Customer_Export_Add" type="Dimension" mode="add" >
      <connection nameref="palodb" />
      <source nameref="Customer_Pipeline" root="All Customers" />
      <dimension name="Customers" />
      </export>

      I am really out of ideas..

      Thanks
      sivgin
    • Still not solved but figured something out.

      For my Customer Dimension I use a Tree (simpliest way and no need of TreePC or TreeNC).

      I tested the load_hierarchy.xml found here in palo-forum.
      This example is a TreePC and with new levels the attributes are correctly mapped. After changing the TreePC into a Tree in load_hierarchy.xml the same error is back:

      PIPELINE CAUSES ERROR:
      <pipeline name="hierarchy_pipe" type="Tree">
      <comment> Pipeline to build up a tree.</comment>
      <sources>
      <source nameref="hierarchy" />
      </sources>
      <target>
      <levels>
      <level>
      <input nameref="element" />
      <attributes>
      <attribute name="english">
      <input nameref="english" />
      </attribute>
      <attribute name="deutsch">
      <input nameref="deutsch" />
      </attribute>
      </attributes>
      </level>
      </levels>
      </target>
      </pipeline>

      RESULT:
      lbl engl. ger.
      =========================
      ZW Zimbabwe Zimbabwe
      no name Zimbabwe Zimbabwe
      (the new entry uses the attributes from the last entry)


      SHOULD BE:
      lbl engl. ger.
      =========================
      no name no name kein Name


      Is anybody using Tree with attributes? Why is that simple example not working?

      Thanks for helping
      sivgin

      The post was edited 1 time, last by sivgin ().

    • Hello

      Try to display the pipeline result using option -d to get an idea.
      If still no clue, try to minimize the data you have to just one or two lines to see the error exactly.
      If still no clue, post the data and the pipeline syntax in the forum so we can check it.

      Regards,
    • I tried also importBiker but it is not working too..
      it is only working when I reimport the project:
      etlclient -s 10.0.50.252 -a ./importBiker/importBiker
      and then start the job
      etlclient -s 10.0.50.252 -p importBiker

      Iam really confused.. why is it working for you?
      Anything special you did adding a new element to a dimension?
      Anything special configureable in ETL Server?

      Thanks for helping
      sivgin
    • Hi Kais,

      To summarize:
      - I do have Palo 3 newest version
      - with parameter d on etlclients output we can see old and the new elements
      - the pipeline is not showing the attributes.. but that is what your documentation is saying
      - olapserver is always running (linux)
      - I start and stop etlserver every day (running hole day) (linux)
      - I assume that my syntax is correct
      - I assume that a standard add mode in my export should add elements, attributes and consolidations

      Can you please tell me that you really tried to make the hirarchy dimension longer?
      I there any possibility that ETL or PALO is caching old values?

      Thanks.. and sorry about that huge Thread with that simple problem
      sivgin
    • Hi Kais,

      today I tried a lot..
      - switch from latest palo to 2.5 back
      - use a new download from etlserver 1.3 and install it. (changes only in etlserver.sh to point to jdk1.6.0_12)
      - etlclient without ssl
      - switching from jdk1.6.0_12 to jdk1.6.0_0
      - logging in etlserver and olapserver (no errors)
      - using alias instead of attribute in project file
      - use validation parameter using etlclient

      All results always with the same error!

      How it works:
      (1)Add my project file (still the same, no changes) to etlserver again and then add new elements usingetlclient.
      (2)Restart etlserver and add new elements afterwards with etlclient.

      Shouldn't that help to understand this issue? I do not believe that OLAP Server is doing something wrong. Should be ETL Server.

      Bizarre is also that the pipeline type TreePC works great without restarting etlserver or reading projectfile (add new elements.. see correct results on Excel add--> fine, cool), unfortunately not Tree!

      I think I did all possible as a user of palo. Would be great hearing from you or the palo team..
      sivgin

      The post was edited 1 time, last by sivgin ().