Batch Apply is the default setting because it is much faster than Transactional Apply.
Transactional Apply mode is generally slower than Batch apply mode because the task will replay the source TLog transactions exactly as they are ordered in the source TLog. Each update to a record will be applied to the target one statement at a time.
Batch Apply mode will do many optimizations to the transactions that it reads from the source Tlog. A couple of examples would be the combining of several statements on one table into a single statement. Like the Insert followed by a couple of update statements would be combined into a single Insert Statement, or several Update statements that would be combined. The most import thing that batch will do is apply all the records in the bac=tch as a single command to the target instead of hundreds, if not thousands of individual statements.
In general batch apply should be used since it is faster.
In some task environments like a Bi Directional task pair it is required to use transactional mode.
Le me know if this helps out and or please accept this as a solution
In addition to the previous comments .In general batch apply should be used since it is faster.
In some task environments like a Bi Directional task pair it is required to use transactional mode or where customer want data in the way it was applied on the Source database.
Where the Changes are very less. in Such case Qlik replicate may wait to create a batch to apply onto target endpoint.
You mentioned "Transactional Apply mode is generally slower than Batch apply mode because the task will replay the source TLog transactions exactly as they are ordered in the source TLog. Each update to a record will be applied to the target one statement at a time."
I have a question about transactional apply mode. So basically, if I ran one update statement on the source, for example, update all records where height > 100cm, which resulted in 10 records being updated on the source. Does this mean that on the target, there will be 10 update statements, rather than one.
Yes, you’re right — in this mode, all the changed rows are applied to the target table one by one, in order, rather than as a single DML operation. This is because, in the transaction log, the 10 UPDATEs are recorded as 10 individual DML operations, even though they were triggered by a single SQL statement.
Hope this helps.
John.
Help users find answers! Do not forget to mark a solution that worked for you! If already marked, give it a thumbs up!
Thanks so much for the quick confirmation! I have another question. Does the change processing mode also affect how changes are applied to the change tables? e.g. if i have transactional apply and change tables enabled, will the records be inserted into the change table one by one as well?
