Issue With SAP Long Text Extraction

Answered

Comments

10 comments

  • Best answer
    Avatar
    Andrew Lund

    Hello Venkata

    For the duplicate long text extraction issue, I advise that you create a support ticket so this can be investigated.

    Regarding other options for extraction long texts, in the past, I have extracted the STXH table for a particular TDOBJECT and then split the TDNAME into it's parts e.g. VBBK is probably a concatenation of the Order Number and Line Item and then based upon the specific orders that I want to extract, I have based then into an Integrate process that is calling the RFC_READ_TEXT / /BOA/ZRFC_READ_TEXT functions. This process then ONLY extracts the extract records you require. From recollections, setting this up in Integrate took 15 to 20 mins and it worked very well.

    I'm happy to help if you need any guidance - just let me know.

     

    Comment actions Permalink
  • Avatar
    Venkata Ravi Kumar Mullamuri

    Hi All,

     

    Adding to that... I didn't face any issue with few other objects like Material, LFA1 etc. But encountering this issue with delivery signatures (TDOBJECT- VBBK) long text, not sure what's the root cause  of it.

    Currently we are in mid of our mock data loads and this issue is delaying our deliverable's.

    Can anyone help me how to fix this issue?

    Thanks & Regards

    Ravi

    0
    Comment actions Permalink
  • Avatar
    Joseph Flesche

    With the project that I work on, we are not extracting that long text object.

    But from past experiences with PLPO, we have seen duplicate long text records extracted. Not many times, but we did see it. We never got down to the root cause, but we identify them in the following manner.

    ;WITH c AS (
    SELECT COUNT(*) OVER (PARTITION BY MANDT, TDOBJECT, TDNAME, TDID, TDSPRAS, [COUNTER] ) AS [Count], *
    FROM TEXT_NAME
    )
    SELECT *
    FROM c
    WHERE [Count]>1

     

    Have you tried to cancel the extract and try again?

    From my experiences, PLPO is the largest text table that we have, we have just shy of 3,000,000 records and it extracts in about 15 minutes from our production system. How many records do you expect to extract for VBBK?

    What is the concern with the duplicate records in the extract? Is it causing load report issues? Have you verified from the front end that the lines are not true duplicates?

    For the duplicate records, is the TDFORMAT and TDLINE fields the same? If so, have you considered just deleting the duplicate records?

    ;WITH c AS (
    SELECT
    COUNT(*) OVER (PARTITION BY MANDT, TDOBJECT, TDNAME, TDID, TDSPRAS, [COUNTER] ) AS [Count],
    ROW_NUMBER() OVER (PARTITION BY MANDT, TDOBJECT, TDNAME, TDID, TDSPRAS, [COUNTER] ORDER BY SequenceNo) AS RowNumber,
    *
    FROM TEXT_NAME
    )
    DELETE FROM c
    WHERE RowNumber>1

    0
    Comment actions Permalink
  • Avatar
    Venkata Ravi Kumar Mullamuri

    Hi Joseph,

    Thanks for your response.

    We have total of 155 million entries in STXL for VBBK and out of which i am trying to extract 20million related to delivery signatures.

    i tried to delete and re triggered the extraction multiple times and the same issue is happening every time.

    As you said, we can remove duplicate entries manually on SQL side once we have data. But so far what i observed is that it extracted same entry five times as of now and extraction job is still active.

    my main concern is whether this extraction job fill finish or will run in infinity loop.

    Also wanted to check is there any workaround to fix this issue.

    Thanks & Regards

    Ravi

    0
    Comment actions Permalink
  • Avatar
    Joseph Flesche

    We have total of 155 million entries in STXL for VBBK and out of which i am trying to extract 20million related to delivery signatures.

    That's a significant amount of records!

    my main concern is whether this extraction job fill finish or will run in infinity loop.

    Agreed.

    Also wanted to check is there any workaround to fix this issue.

    The only thing that I would suggest trying at this point is to see if entering a specific TDNAME (one that is known to be duplicated) in the DSP configuration to see if it is duplicated.

     

    0
    Comment actions Permalink
  • Avatar
    Venkata Ravi Kumar Mullamuri

    Hi Joseph,

    When i enter specific TDNAME and extract, i am not facing any duplicate issue. This issue is occurring only when i try to extract all 20million records at once.

    Is there any possibility to pass multiple values in TDNAME at once?

    Thanks & Regards

    Ravi

    0
    Comment actions Permalink
  • Avatar
    Joseph Flesche

    Is there any possibility to pass multiple values in TDNAME at once?

    I have asked this question in the past. The answer I received is no, it is not possible. How many TDNAME values do you need to extract?

    0
    Comment actions Permalink
  • Avatar
    Venkata Ravi Kumar Mullamuri

    Hi Joseph,

    it will be in few lakhs.

    Thanks & Regards

    Ravi

    0
    Comment actions Permalink
  • Avatar
    Venkata Ravi Kumar Mullamuri

    Hi Andrew,

    Thanks for your response. I will raise a support ticket for this duplicate issue.

    Can you please share any links/videos related to Integrate as i am new to Integrate functionality.

    Thanks & Regards

    Ravi

    0
    Comment actions Permalink
  • Avatar
    Joseph Flesche

    Andrew Lund,

    I am also interested to understand how this works.

    0
    Comment actions Permalink

Please sign in to leave a comment.