Topic
  • 2 replies
  • Latest Post - ‏2013-05-16T08:04:41Z by VMNanduri
VMNanduri
VMNanduri
13 Posts

Pinned topic WTX vs FFD- Which one is suggestable for large volume of data transformations(CSV to XML)?

‏2013-05-16T07:13:37Z |

Hi ,

We have a requirement to transform big CSV data contains 600 fields[columns] and 100 records to XML format. We knew that we can implement this in datapower using either WTX or FFD. But we doubt which approach would be better in view of performance and other key factors .

Can you suggest which one (WTX or FFD) is more feasible in view of performance ? If possible can you also advice pros and cons of using WTX/FFD for this kind of big transformation?

Thanks in Advance!!

Thanks,

Lakshman.

 

  • HermannSW
    HermannSW
    5921 Posts

    Re: WTX vs FFD- Which one is suggestable for large volume of data transformations(CSV to XML)?

    ‏2013-05-16T07:56:29Z  

    As always, it depends -- you should do performance testing to compare which is best for your scenario .

    Basically you have three options:

    1. WTX
    2. Contivo Analyst (generates XSLT(s) and FFD(s), needs a product license,
      http://www-01.ibm.com/support/docview.wss?uid=swg21366685)
    3. one of the simple supported FFDs
      (http://www-01.ibm.com/support/docview.wss?uid=swg27022977)


    It is no option to handcraft a FFD for the whole task, that is not supported and any PMR you will raise will be closed.

    You can find a sample implementation for (3) in this Blog posting:
    https://www.ibm.com/developerworks/community/blogs/HermannSW/entry/csv2html_xsl13

    But you have to deal with how to handle cases where $colsep appears (quoted) in fields.
    If you do not want to do that, probably (1) and (2) are better options.


    Hermann<myXsltBlog/> <myXsltTweets/> <myCE/>

     

  • VMNanduri
    VMNanduri
    13 Posts

    Re: WTX vs FFD- Which one is suggestable for large volume of data transformations(CSV to XML)?

    ‏2013-05-16T08:04:41Z  
    • HermannSW
    • ‏2013-05-16T07:56:29Z

    As always, it depends -- you should do performance testing to compare which is best for your scenario .

    Basically you have three options:

    1. WTX
    2. Contivo Analyst (generates XSLT(s) and FFD(s), needs a product license,
      http://www-01.ibm.com/support/docview.wss?uid=swg21366685)
    3. one of the simple supported FFDs
      (http://www-01.ibm.com/support/docview.wss?uid=swg27022977)


    It is no option to handcraft a FFD for the whole task, that is not supported and any PMR you will raise will be closed.

    You can find a sample implementation for (3) in this Blog posting:
    https://www.ibm.com/developerworks/community/blogs/HermannSW/entry/csv2html_xsl13

    But you have to deal with how to handle cases where $colsep appears (quoted) in fields.
    If you do not want to do that, probably (1) and (2) are better options.


    Hermann<myXsltBlog/> <myXsltTweets/> <myCE/>

     

    Hi Hermann,

    Thanks for your input. I hope this helps us to predict better approach before we start implementation.

    Thanks,

    Lakshman.