IBM Support

JR30649: READERS-PER-NODE AND FILTER OPTION ON SEQUENTIAL STAGE SHOULD BE MUTUALLY EXCLUSIVE

Subscribe

You can track all active APARs for this component.

 

APAR status

  • Closed as program error.

Error description

  • The filter option should really be made mutually exclusive with
    multiple readers as in most situations it will not function as
    desired.
    
    The problem will occur with any filter that causes the number of
    bytes read to be different (in this case less) than the number
    of bytes in the record.  The reason for this is the way that the
    multiple readers functionality works.  In the case of a
    delimited file it is split up based on its total byte length.
    Taking a simple case where the file length is 256 and we have 2
    readers, reader 1 will be set to read from byte 0 to 128 and
    reader 2 from 128 to 256.  Now it is very likely byte 128 falls
    in the middle of a record.  In this case reader 1 will read from
    0 to 128 plus the number of bytes to the next record delim.
    Reader 2 will start reading at byte 128 until it finds the first
    record delim and ignore this data, starting actual reading at
    the next record.  This guarantees there is no overlap and the
    correct number of records is read.
    Sequential file stage using multiple readers and filter option:
    example:
    -readers 2
    -filter 'cut -c1-4'
    
    The problem occurs when the first reader is calculating its
    current offset.  Since the filter is passing fewer bytes than
    are actually in the file, the first reader will calculate its
    offset to be earlier in the file than it actually is meaning it
    will read a bunch of records past where it is supposed to stop.
    
    Unfortunately the solution to this would mean preprocessing the
    file which would negate any performance improvements of running
    multiple readers on the file.
    
    The answer to this is this combination of
    arguments is not supported and should be flagged as such.
    

Local fix

  • do not use filter and multiple node option together
    

Problem summary

  • ****************************************************************
    USERS AFFECTED:
    All versions of IS
    ****************************************************************
    PROBLEM DESCRIPTION:
    When the filter and readers options are used in conjunction the
    results are only correct under certain situations.  The user is
    not made aware of that.
    ****************************************************************
    RECOMMENDATION:
    Install patch JR30649
    This change is included in 8.1 Fix Pack 1.
    ****************************************************************
    

Problem conclusion

  • A warning was added to alert the user to the possibility of
    improper function in 8.0.1 and beyond.
    

Temporary fix

Comments

APAR Information

  • APAR number

    JR30649

  • Reported component name

    WIS DATASTAGE

  • Reported component ID

    5724Q36DS

  • Reported release

    752

  • Status

    CLOSED PER

  • PE

    NoPE

  • HIPER

    NoHIPER

  • Special Attention

    NoSpecatt

  • Submitted date

    2008-10-02

  • Closed date

    2009-01-20

  • Last modified date

    2009-12-17

  • APAR is sysrouted FROM one or more of the following:

  • APAR is sysrouted TO one or more of the following:

Modules/Macros

  • SERVER
    

Fix information

  • Fixed component name

    WIS DATASTAGE

  • Fixed component ID

    5724Q36DS

Applicable component levels

  • R752 PSN

       UP

[{"Business Unit":{"code":"BU059","label":"IBM Software w\/o TPS"},"Product":{"code":"SSVSEF","label":"IBM InfoSphere DataStage"},"Component":"","ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"7.5.2","Edition":"","Line of Business":{"code":"LOB10","label":"Data and AI"}}]

Document Information

Modified date:
17 December 2009