SingleCommit with MultipleCommit

[Niko]

Anyone know what is the difference between SingleCommit with MultipleCommit When i set a chore task?

[Wim Gielis]

The difference is a commit step after each of the TIs, or only 1 commit after all TIs are finished.

[Alan Kirk]

The difference is a commit step after each of the TIs, or only 1 commit after all TIs are finished.

With one potential "gotcha" proviso; it only applies to TI processes which are called from the chore itself. If you have to call (say) an upload process a variable number of times, and therefore use a loop to run a series of ExecuteProcess() functions, the commit will not happen until after the master process that makes the calls has finished, regardless of whether single or multi commit is chosen.

[Niko]

The difference is a commit step after each of the TIs, or only 1 commit after all TIs are finished.

With one potential "gotcha" proviso; it only applies to TI processes which are called from the chore itself. If you have to call (say) an upload process a variable number of times, and therefore use a loop to run a series of ExecuteProcess() functions, the commit will not happen until after the master process that makes the calls has finished, regardless of whether single or multi commit is chosen.

thanks for your reply~
Now suppose i created a chore called "Chore A", this chore calls two processes, the first process is called P1, another is P2,
P1 Transfer data from a csv file to CUBE A, P2 Transfer data from CUBE A to CUBE B
Is there a difference between singlecommit and multiplecommit in this situation?

[declanr]

The difference is a commit step after each of the TIs, or only 1 commit after all TIs are finished.

With one potential "gotcha" proviso; it only applies to TI processes which are called from the chore itself. If you have to call (say) an upload process a variable number of times, and therefore use a loop to run a series of ExecuteProcess() functions, the commit will not happen until after the master process that makes the calls has finished, regardless of whether single or multi commit is chosen.

thanks for your reply~
Now suppose i created a chore called "Chore A", this chore calls two processes, the first process is called P1, another is P2,
P1 Transfer data from a csv file to CUBE A, P2 Transfer data from CUBE A to CUBE B
Is there a difference between singlecommit and multiplecommit in this situation?

In a multi commit scenario the data loaded to Cube A is available to everyone immediately when P1 is finished. So if a user queries Cube A while P2 is still running they can see the data and not hold up P2.
In single commit mode; the chore will hold a lock on Cube A so the user won’t be able to see the cube data results until the whole chore finishes. (This is assuming the user tries to read data after P1 is finished but before P2 finishes.)

[Niko]

With one potential "gotcha" proviso; it only applies to TI processes which are called from the chore itself. If you have to call (say) an upload process a variable number of times, and therefore use a loop to run a series of ExecuteProcess() functions, the commit will not happen until after the master process that makes the calls has finished, regardless of whether single or multi commit is chosen.

thanks for your reply~
Now suppose i created a chore called "Chore A", this chore calls two processes, the first process is called P1, another is P2,
P1 Transfer data from a csv file to CUBE A, P2 Transfer data from CUBE A to CUBE B
Is there a difference between singlecommit and multiplecommit in this situation?

In a multi commit scenario the data loaded to Cube A is available to everyone immediately when P1 is finished. So if a user queries Cube A while P2 is still running they can see the data and not hold up P2.
In single commit mode; the chore will hold a lock on Cube A so the user won’t be able to see the cube data results until the whole chore finishes. (This is assuming the user tries to read data after P1 is finished but before P2 finishes.)

i understand , thank u.