Multi-Threading and Parallelism
Due to the high data volumes of most fact data (such as sales and inventory), it is necessary to process the data using
multiple CPU threads on the database. In RAP’s second-generation architecture, multi-threading is handled automatically. You
must not attempt to alter any threading parameters to force a specific thread count greater than 1. If you are providing any
data files ending in a .dat
extension, then it might contain an ETL_THREAD_VAL
column. This
column must be hard-coded to be 1
on all rows without exception; it should not be null and should not be
any value greater than 1. Similarly, there are database parameters named LOC_NUM_OF_THREAD
in the C_ODI_PARAM_VW
table. These must be set to a value of 1
and should not be altered to any value
greater than 1.