Migration Tablesplitting: Faq about Creation of Temporary Indexes

[Former Member]

Dear Experts,

We are in the Process of Test migration of our ECC System  with SolarisSpacrc5.10 & Oracle11.3 to RHEL6.4 with MAXDB7.8

For optimization we have done Table splitting on some of the Large Tables including cluster tables with standard R3ta table splitting by using SWPM.

After Table splitting it has generated WHR files and then we moved this WHR files and whr.txt file to EXPORT/ABAP/DATA directory and also from the table splitting installation directory TABLE_IDX.STR files also we moved to EXPORT/ABAP/DATA directory and then started normal DB Export, and in the middle during R3SZCHK it is throwing below errors:

more R3szchkExport.log

ERROR: add_fld: index field BELNR (index FAGL_SPLINFO~IMG) not in previous table

ERROR: add_fld: index field BELNR (index FAGL_SPLINFO~IMG) not in previous table

ERROR: add_fld: index field PERNR (index PPOIX~IMG) not in previous table

ERROR: add_fld: index field PROGNAME (index DYNPSOURCE~IMG) not in previous table

ERROR: add_fld: index field AWREF (index ACCTIT~IMG) not in previous table

ERROR: add_fld: index field BELNR (index FAGL_SPLINFO_VAL~IMG) not in previous table

The above error is Due to temporary indexes added after table splitting to the STR files, so in this case do we need to create the temporary indexes as per TABLE_IDX.STR files . AS per System Copy guide it is recommended to create the temporary indexes on Oracle DB based on R3ta_hint file for the Tables to be split..and after Export we can drop those indexes as per guide

there is some procedure to create the temporary indexes by using R3load in the guide, so can we create the temporary indexes for cluster tables also which we have selected for Split-Please suggest

But temporarily for the export , I have deleted the TABLE_IDX.STR files from the directory EXPORT/ABAP/DATA and then resumed the Export then it has created the STR files for the Splitted tables and Multiple packages also has been processed for the splitted tables along with multiple TOC files as per splitted tables. So here in this scenario export has been completed successfully with automatic STR file created by tool and multiple packages has been created and exported for splitted tables. The downtime for the Export also has been reduced compared to previous run.

But my doubt is can we Import the same Export which we have taken into the Target DB or else as per guide do we need to create the Temporary indexes and then add the Table_IDX.STR files to EXPORT/ABAP/DATA directory and then DB Export should be performed?

Experts Please suggest in this case how can we proceed, as I am new to Table Split Procedure , I have seen some thaeads regarding this but not found any solution-Please Help

Regards,

Samba

[CharlesChen]

Hi Samba,

To be frankly, I don't know whether any issue will happen during the import.

But here are what I know.

1.  Creating temporary indices will help to improve the performance.

And the below are mentioned as prerequisites in the guide.

**********

For the following databases, we recommend that you create an index on the column listed in the hints file for the table you want to split before starting R3ta:

IBM DB2 for z/OS (Only create a temporary index if you want to perform an unsorted unload.)

Oracle

**********

2. Normally, deleting files generated by tools without guidelines is not a good idea. It often leads to unexpected issues.

Best regards,

Charles