IBM Support

PI85456: ANALYZE FAILS ON BIG SQL PARTITIONED TABLES THAT HAVE A MIX OF DATE AND TIMESTAMP DIRECTORY STRUCTURES ON HDFS

Subscribe

You can track all active APARs for this component.

 

APAR status

  • Closed as program error.

Error description

  • For partitioned tables, there is a directory on HDFS for every
    partitioning value.  Big SQL expects a certain format on HDFS
    for date values when a table is partitioned on a Date type.
    When the format on HDFS deviates from the format that Big SQL
    is expecting, Analyze may fail.
    
    For example, when a Big SQL table is partitioned by a date
    type, by default it is stored as a timestamp type in Hive.
    When a Big SQL table is partitioned by a date stored as date
    type it is stored as a date type in Hive.  When a Hive table is
    imported into Big SQL with a Date type then Big SQL expects
    values of Date type.
    
    If the directory structure contains a mix of Date and Timestamp
    values then Big SQL Analyze can get confused.
    

Local fix

  • -Create a new Big SQL table (tab2) partitioned by Date in a new
    location on HDFS.  Big SQL will be expecting timestamp
    formatted directory structures on HDFS
    -Copy all timestamp formatted directories from the old table
    (tab1) to the new table on HDFS
    -Copy all date formatted directories from the old table to the
    new table on HDFS renaming the date directories to timestamp
    directories; e.g.:
           hadoop fs -cp /user/hadoop/tab1/c2=2017-07-15
    /user/hadoop/tab2/c2= 2017-07-15 00%3A00%3A00
    -Issue an hcat-sync command: CALL SYSHADOOP.HCAT_CACHE_SYNC
    ('bigsql','tab2');
    -Run the Analyze command on tab2
    Change applications to use tab2 instead of tab1
    

Problem summary

  • See error description
    

Problem conclusion

  • The problem is fixed in Version 5.0.1.0 and later fix packs
    

Temporary fix

Comments

APAR Information

  • APAR number

    PI85456

  • Reported component name

    INFO BIGINSIGHT

  • Reported component ID

    5725C0900

  • Reported release

    425

  • Status

    CLOSED PER

  • PE

    NoPE

  • HIPER

    NoHIPER

  • Special Attention

    NoSpecatt / Xsystem

  • Submitted date

    2017-08-03

  • Closed date

    2018-03-01

  • Last modified date

    2018-03-01

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

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

Modules/Macros

  • Unknown
    

Fix information

  • Fixed component name

    INFO BIGINSIGHT

  • Fixed component ID

    5725C0900

Applicable component levels

  • R425 PSY

       UP

[{"Business Unit":{"code":"BU059","label":"IBM Software w\/o TPS"},"Product":{"code":"SSCRJT","label":"IBM Db2 Big SQL"},"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"425","Line of Business":{"code":"LOB10","label":"Data and AI"}}]

Document Information

Modified date:
25 August 2020