Hi, I am experiencing an issue and it looks like there is some limit on the number of DBSW formulas when sending data back to a cube.
The file originally works fine for writing data back in Perspectives and it takes around 10 minutes to complete.
When I put it with PAfE v55, it refreshed but not sending any values to the cube at all no matter what I tried(F9,Shift+F9,Crtl+Alt+F9,Ctrl+Alt+Shift+F9, etc). I reduced the number of DBSW cells (just deleting lines) back to around 220,000 cells and it worked. 
I would expect better performance in PAfE without changing anything in the original file and didn't have any impression that IBM put any limits on that?

------------------------------
mvp morgan
------------------------------

#PlanningAnalyticswithWatson

As a possible workaround have you thought about using a quick report and formulas linking to your values to load back?  Then you can use the commit function when ready to send.

Not sure on why it would fail at 220k cells other than maybe a timeout?  Have you run a network trace? Or enabled debugging to see what's going on?

------------------------------
Robby Meyers
------------------------------

Thanks, Robby. I didn't see any error message and I could see some REST API statements were running from TM1Top. The network is fine at that time. The sheet was refreshed but showing all 0 values, which shouldn't and the cube was showing nothing.
Yes, a quick report could be solution but just wondering why there are difference which I don't expect.

------------------------------
mvp morgan
------------------------------

Original Message:
Sent: Wed January 27, 2021 12:29 AM
From: Robby Meyers
Subject: Cell Number Limit for DBSW?

As a possible workaround have you thought about using a quick report and formulas linking to your values to load back?  Then you can use the commit function when ready to send.

Not sure on why it would fail at 220k cells other than maybe a timeout?  Have you run a network trace? Or enabled debugging to see what's going on?

------------------------------
Robby Meyers
------------------------------

What symptoms are you seeing of the failure?
With a DBS we do see the formula fail to display a cell value but the data has still been updated in TM1.  It seems that the second pass to retrieve fails but the send works.

One other thing to watch out for with DBS is that if you don't have write access for some reason then the formula will return the cell value rather than an error (like in the legact client) this can make it harder to spot an issue.

------------------------------
Steven Rowe
------------------------------

Thanks, Steven. As I said above, I din't see any symptoms of failure and the sheet are showing all 0 values which are supposed to be the value from other calcualted sheets. Nothing has been sent to the cube after several minutes of refreshing in the Excel. I am using my admin account to test it and it works fine with less lines.

------------------------------
mvp morgan
------------------------------

Original Message:
Sent: Wed January 27, 2021 03:54 AM
From: Steven Rowe
Subject: Cell Number Limit for DBSW?

What symptoms are you seeing of the failure?
With a DBS we do see the formula fail to display a cell value but the data has still been updated in TM1.  It seems that the second pass to retrieve fails but the send works.

One other thing to watch out for with DBS is that if you don't have write access for some reason then the formula will return the cell value rather than an error (like in the legact client) this can make it harder to spot an issue.

------------------------------
Steven Rowe
------------------------------

BTW, they are DBSW rather than DBS.

------------------------------
mvp morgan
------------------------------

Original Message:
Sent: Wed January 27, 2021 03:54 AM
From: Steven Rowe
Subject: Cell Number Limit for DBSW?

What symptoms are you seeing of the failure?
With a DBS we do see the formula fail to display a cell value but the data has still been updated in TM1.  It seems that the second pass to retrieve fails but the send works.

One other thing to watch out for with DBS is that if you don't have write access for some reason then the formula will return the cell value rather than an error (like in the legact client) this can make it harder to spot an issue.

------------------------------
Steven Rowe
------------------------------

what version of TM1 is in use?

------------------------------
Ted Phillips
------------------------------

PAL 2.0.9.1 and PAW v54 and PAfE v55

------------------------------
mvp morgan
------------------------------

Original Message:
Sent: Wed January 27, 2021 04:52 PM
From: Ted Phillips
Subject: Cell Number Limit for DBSW?

what version of TM1 is in use?

------------------------------
Ted Phillips
------------------------------

That TM1 is sufficiently new where I'm not aware of any known capacity issues.  Better to generate a PA for Excel log bundle at 'all' level and pass it along to IBM for review.

------------------------------
Ted Phillips
------------------------------

Original Message:
Sent: Wed January 27, 2021 09:26 PM
From: mvp morgan
Subject: Cell Number Limit for DBSW?

PAL 2.0.9.1 and PAW v54 and PAfE v55

------------------------------
mvp morgan
------------------------------


Original Message:
Sent: Wed January 27, 2021 04:52 PM
From: Ted Phillips
Subject: Cell Number Limit for DBSW?

what version of TM1 is in use?

------------------------------
Ted Phillips

Hi there:

Apologies in advance for asking this question ... but typically when you are moving large amounts of data between TM1 cubes, it is better to write a TurboIntegrator (TI) process to do it so my question is - is there a reason why you are NOT doing that?

Actually, I realize there are quite a few reasons why you might not use a TI including (for example):
- The Excel logic is quite intricate and complicated, and it would be difficult to reverse engineer, or perhaps it changes often
- The person who owns this process is not familiar with TI script, and feels better able to control the logic from Excel with DBSW
- We could but we do this so rarely that it is not worth the effort of developing the TI
- Just don't have time!

These are all perfectly good reasons.

All that said, here are some reasons why you might find value in shifting to TI:
- You can end up with a much more robust process (if done properly)
- The processing time will likely drop from minutes to seconds (which doesn't matter if you do not do this often)

Obviously I don't know all the details of your situation, and I do not mean to second-guess you - but I guess I am just curious. Is it for one of the reasons I mentioned, or for some reason I did not think of?

Many thanks,
AG

------------------------------
Ann-Grete Tan
Chief Strategy & Marketing Officer
QueBIT Consulting

Analyze. Plan. Achieve.
https://quebit.com
------------------------------

Thanks, that's what I am currently working towards rather than battling with potential bugs in Excel.

------------------------------
mvp morgan
------------------------------

Original Message:
Sent: Thu January 28, 2021 05:46 PM
From: Ann-Grete Tan
Subject: Cell Number Limit for DBSW?

Hi there:

Apologies in advance for asking this question ... but typically when you are moving large amounts of data between TM1 cubes, it is better to write a TurboIntegrator (TI) process to do it so my question is - is there a reason why you are NOT doing that?

Actually, I realize there are quite a few reasons why you might not use a TI including (for example):
- The Excel logic is quite intricate and complicated, and it would be difficult to reverse engineer, or perhaps it changes often
- The person who owns this process is not familiar with TI script, and feels better able to control the logic from Excel with DBSW
- We could but we do this so rarely that it is not worth the effort of developing the TI
- Just don't have time!

These are all perfectly good reasons.

All that said, here are some reasons why you might find value in shifting to TI:
- You can end up with a much more robust process (if done properly)
- The processing time will likely drop from minutes to seconds (which doesn't matter if you do not do this often)

Obviously I don't know all the details of your situation, and I do not mean to second-guess you - but I guess I am just curious. Is it for one of the reasons I mentioned, or for some reason I did not think of?

Many thanks,
AG

------------------------------
Ann-Grete Tan
Chief Strategy & Marketing Officer
QueBIT Consulting

Analyze. Plan. Achieve.
https://quebit.com
------------------------------