- For Oracle, run execute SP_POPULATE_PATTERN_LOCK(2), where degree is 2.
- For DB2, run call
SP_POPULATE_PATTERN_LOCK(2), where degree is 2.
- For SQL Server, run EXEC [dbo].[SP_POPULATE_PATTERN_LOCK]
@p_parallel_degree = 2, where degree is 2.
After you run the script, the UARI_PATTERN_LOCK table is populated
with the stored procedures with the degree value. This degree value
is configurable. Increase the degree of parallel execution for the
Interact Event Pattern report aggregation process to reduce the elapse
time. If the degree is set to higher value, hardware resource requirements
also increase proportionally. The number of procedures that are run
for data aggregation depend on the degree value.
At the first
run, ETL does not insert any values against the respective PatternID
in the UARI_DELTA_PATTERNS table, as all the patterns are new or delta.
The reports aggregation process collects all the PatternIDs from the
ETL tables inserts them in the UARI_DELTA_PATTERNS table.
The
ETL process calls the SP_AGGR_RUN_STATUS procedure. The SP_AGGR_RUN_STATUS
procedure checks the UARI_PATTERN_LOCK table for running JobIDs. The
JobID contains the flag 'Y' either in case of running or failed scenarios.
If there are any failed jobs, the procedure sets the flag to 'N'.
For running jobs, the value remains 'Y'. The ETL process always checks
the status of the reports aggregation by checking status of the submitted
jobs. If the ETL finds reports aggregation running, the ETL does not
start its execution. The ETL starts after again as per the schedule.
The ETL process checks the UARI_PATTERN_LOCK table for the number
of JobIDs with value Y. The ETL process starts only
if the number of JobIDs with the value 'Y' is 0. If the number of
JobIDs with value 'Y' is greater than 0, then the ETL process is skipped
and it runs at the next scheduled interval. For more information about
the ETL process, see the IBM® Interact Administrator's Guide.
From the second run
onwards, the ETL process updates the UARI_DELTA_PATTERNS table with
the update flag for the updated PatternID, for updated data, the Pattern
ID is marked with 'U'. For deleted data, the PatternID is marked with
'D'.Newly added Patterns are identified by reports aggregation code
and then inserted into UARI_DELTA_PATTERNS with the flag 'P'. The
ETL always updates the flags in UARI_DELTA_PATTERNS table for the
existing PatternID's and newly added patternID's are processed and
inserted into UARI_DELTA_PATTERNS table by reports aggregation process.
The aggregation process is done for only the PatternIDs marked
with the 'U' or 'D' flag.