Page 1 of 1

CellPutN replacing with zeros

Posted: Fri Aug 23, 2024 10:24 pm
by Nance1234
Hi all,

I have 2 questions:

1) Do we really need a zerout to be done in the TI process while using CellPutN for data load since it replaces the existing numbers?
2) If data is being pulled via SQL and we have used CellputN for load into the cube.
source has numbers -> ran the process -> data got loaded into the cube
Source deleted some numbers -> ran the process -> will it replace the existing numbers with a zero in those target cells.

I think yes, it should replace the existing numbers in the target cells with a zero but need confirmation from experts here.

Thanks a lot in advance.

Re: CellPutN replacing with zeros

Posted: Sat Aug 24, 2024 3:13 pm
by Wim Gielis
Yes typically you do a zero out such that combinations that had values previously, but not at the time of a new load, are empty/do not exist anymore/do not take up space anymore.
A process without zero out is possible too, but are less frequent. It depends on the use case.

Re: CellPutN replacing with zeros

Posted: Mon Aug 26, 2024 8:45 am
by lotsaram
Nance1234 wrote: Fri Aug 23, 2024 10:24 pm source has numbers -> ran the process -> data got loaded into the cube
Source deleted some numbers -> ran the process -> will it replace the existing numbers with a zero in those target cells.
This will depend. If the numbers deleted from the source are represented in the query as records containing zero or null then YES, in this case the values in the TM1cube would reflect the source and tie back. However, if there are no longer any records in the query representing the values which have been deleted then NO, the original values will still be present in the TM1 cube from the first load and the values in TM1 will no longer tie back to the source.

So if you are confident that you will have records containing zero or null for the deleted data then it is safe to skip the ViewZeroOut as the CellPutN on the data tab will correctly fill the cells with zero. But if it could be that the deleted data isn't represented in the query result at all then you must clear the cells in TM1 before loading.

Re: CellPutN replacing with zeros

Posted: Mon Aug 26, 2024 4:37 pm
by Nance1234
Wim Gielis wrote: Sat Aug 24, 2024 3:13 pm Yes typically you do a zero out such that combinations that had values previously, but not at the time of a new load, are empty/do not exist anymore/do not take up space anymore.
A process without zero out is possible too, but are less frequent. It depends on the use case.
Thank you, Wim Gielis :)

Re: CellPutN replacing with zeros

Posted: Mon Aug 26, 2024 4:44 pm
by Nance1234
lotsaram wrote: Mon Aug 26, 2024 8:45 am
Nance1234 wrote: Fri Aug 23, 2024 10:24 pm source has numbers -> ran the process -> data got loaded into the cube
Source deleted some numbers -> ran the process -> will it replace the existing numbers with a zero in those target cells.
This will depend. If the numbers deleted from the source are represented in the query as records containing zero or null then YES, in this case the values in the TM1cube would reflect the source and tie back. However, if there are no longer any records in the query representing the values which have been deleted then NO, the original values will still be present in the TM1 cube from the first load and the values in TM1 will no longer tie back to the source.

So if you are confident that you will have records containing zero or null for the deleted data then it is safe to skip the ViewZeroOut as the CellPutN on the data tab will correctly fill the cells with zero. But if it could be that the deleted data isn't represented in the query result at all then you must clear the cells in TM1 before loading.
Thank you Lotsaram for this detailed explanation.
I had no issues adding a zerout but was still confused that when it's CellPUT then why the data at the source not matching in the cube. Your explanation made it clear, there must be some deleted records not represented in the query result which is why the numbers for the target cells in the cube remain populated even though the values were zero at the source.
Thanks again :)