1.Create the Environment Section of the script a.JOBNAME – also referred to as the Apply Engine Name, uniquely identifies the Apply Engine in the parsed script and is normally the "Subscriber name" passed by the Apply Engine when requesting connection to a Capture/Publisher. b.COMMIT – Only required if the Apply Engine is not processing normal published CDC data that automatically identify Unit-of-Work (UOW) commits. c.REPORT – optional, specifies that a progress report will be produced after "n" (number) of Source records processed. 2.Create the Datastore Definition Section of the script. a.RDBMS – required for relational datastores. b.DESCRIPTION – derived from the datastore layouts in steps 1b and 2b. c.DATASTORE – derived from the physical objects in steps 1a and 2a. 3.Create the mapping procedures based on the information obtained in steps 3 and 4 unless it was determined in step 3a that this a "Pure Replication" script. 4.Create any Cursors that perform lookups and or joins based on the information obtained in step 4. 5.Populate the Field Specification Section based on any special processing requirements found in step 4. This can include date formatting (DATEFORMAT), redefining a source field (EXTTYPE), setting an invalid source field value to a default value (INVALID), etc. 6.Create the Processing Section of the script. a.Determine the target update strategy and use the appropriate target operation command (i.e. INSERT or CHANGE.) b.Reference the target datastores defined in step 5b after the target operation command. c.Reference the mapping procedures, if used, within the SELECT/FROM command. If this was determined to be a "Pure Replication" script, specify the REPLICATE command. Note, in extremely simple scripts that do require mapping, the source to target mapping can be specified ‘inline’ between the SELECT and the FROM keywords. d.Reference the source datastore derived in step 5b after the FROM keyword. |
