Runtime Analysis for DTPs/Transformations

[sascha_mink2]

Ever wondered where your DTP/Transformation spent it's time?

This is how you can find out:

At first you have to create a small programm like the following:

REPORT ZTEST.

DATA: R_DTP TYPE REF TO CL_RSBK_DTP.

                " loading the dtp

CALL METHOD CL_RSBK_DTP=>FACTORY

    EXPORTING

         " here you have to place the technical name of your dtp

         I_DTP  = 'DTP_4YIEJ....'

    RECEIVING

         R_R_DTP = R_DTP.

DATA L_R_REQUEST TYPE REF TO CL_RSBK_REQUEST.

L_R_REQUEST = R_DTP->CREATE_REQUEST( ).

" we have to run it in sync-mode (not in Batch)

L_R_REQUEST->SET_CTYPE( RSBC_C_CTYPE-SYNC ).

L_R_REQUEST->DOIT( ).

Now run this programm through Transaction SE30 or SAT :

After the execution of the DTP you will receive the result. In our case it was an select on the psp element. We created an index and the dtp needed only 2 minutes instead of 30 minutes.

Of course you can also simply call the transaction RSA1 through SAT (using the processing mode "serially in the dialog process (for debugging)"). But doing it that way you have to filter out the overhead in the performance log created by rsa1 and the data will not be written to the target (it's only simulated) so you might miss sometimes a bottleneck in your dtp/Transformation.

thanks for reading ...