Skip to Content
0
Former Member
Aug 05, 2015 at 07:21 PM

SPLITTED_LOAD option in r3load

395 Views

Hi All,

Has anyone tried the option SPLITTED_LOAD for importing table data parallely?

This is the documentation which i can see in the SAP note.

  • SPLITTED_LOAD ( SAP Kernel Version 7.20 and higher)
    This is an expert
    option to be able to use DB2 LOAD in parallel for a gradually exported table.
    Prerequisites:

    • The table was divided using WHERE clauses during the export and is loaded
      with multiple R3load processes.

      • The target table does not have indexes yet.

      Each R3load process creates a copy of the target
      table during import and uses DB2 LOAD to load this copy. Before creating the
      first index on the target table, all data is copied from the table copies to the
      target table using DB2 LOAD FROM CURSOR. If you use the option SPLITTED_LOAD,
      you can make time when using the duplicate loading process: LOAD FROM CURSOR is
      significantly faster than LOAD via the CLI client, and the loading of table
      copies can happen in parallel. You can use the environment variable
      DB6LOAD_TEMP_TBSPACE to specify, in which table space the table copies are
      created.

      A large LOAD FROM CURSOR ... SELECT statement is
      used to copy the data from the table copies to the target table. The SELECT part
      of this statement can be accelerated by activating SMP parallelism (SAP Note 2047006). To do so, it is
      sufficient to set INTRA_PARALLEL=YES and MAX_QUERYDEGREE='ANY'. You can leave
      the value for DFT_DEGREE as 1.