The LAST_UPD column in Siebel transactional database tables is used for incremental change capture. This timestamp reflects the actual event time. It is therefore possible for a data row to be committed to the transactional database with a LAST_UPD date that is older than the date on which the last refresh was executed. This will cause the data row to be missed in the subsequent extract (if based purely on LAST_UPD date). However, the LAST_UPD date column still provides an opportunity to improve the change capture process by overlapping the extraction date window by the number of days set in this parameter. The records extracted in the overlapped window are filtered by comparing this information with information in the Image table. The Prune Days setting ensures that the rows that had values in LAST_UPD older than values in LAST_REFRESH_DATE are not missed. This is a parameter that can be set based on experience with processes, such as remote sync, that potentially can cause records to be missed. This parameter cannot be less than 1. For example: Assume the table W_PERSON_D was refreshed on January 15th by querying the table S_CONTACT. And, the Prune Days setting was set to 5. The next time S_CONTACT is queried to load W_PERSON_D, the records that have a LAST_UPD value since January 10 are compared with the ROW_ID of the Image table to cover for any missing records between January 15 and January 10 (the overlap period). |