PARALLEL UPDATE STATISTICS

BRCONNECT supports parallel processing with:
 Parallel threads, especially for large numbers of tables
 The DBMS_STATS package, especially for large tables
Parallel threads, if activated, significantly improves run times. You can implement this as follows:
 Command call brconnect -p <number of threads>
 Parameter stats_parallel_degree in the init<SID>.sap file
Here are some examples:
brconnect -u / -c -f stats –t all -p 4
stats_parallel_degree = 4
Each of these sets the number of threads – that is, the degree of parallelism – to 4. The default is 1, which
means that update statistics runs in non-parallel – that is, serial – mode. In either case, a single thread
processes each table.
For more information on parallel processing with BRCONNECT, see SAP Note 403713.
You can also use the DBMS_STATS package to update statistics in parallel for individual tables. The 
parallel processing option in this package considerably reduces run times for very large tables (that is,
several GB and more). Each table can be processed by several threads.
For more information on parallel processing with DBMS_STATS, see SAP Note 408532.
You can use both the above types of parallel processing in the same BRCONNECT run