Does the introduction of a rule to an area ever remove any existing data in that area?

[JohnO]

Well there's more to it than the title.

We are using 2.0.19. For our depreciation model we have been loading forecast depreciation from an external system. We recently changed that to TM1 calculating the forecast. Before deploying the rules I did not clear all of the existing forecast data. These rules have been in place for about a month with daily saves and several weekly server restarts and everything appeared to be working fine. The other day we altered the rules so that forecast depreciation for one account was no longer being rule defined and the old data appeared.

So at this stage the question is whether we should expect TM1 to hold the underlying data forever? I have other rule defined data where users make selections which drive whether they can enter data for particular month/ account combinations and if they cannot then the data values are set to 0 by the rules so I'm wondering if any pre-existing data is kept forever in the cases where rules are invoked and the rule sets the area to 0.

Continuing with the story, for context our depreciation cube is linked to our P&L cube via rules and the forecast numbers should always match between them with multiple accounts involved. Back to the case above I cleared out the data that had reappeared for the account in question and the P&L cube matched back to the depreciation cube for that account but then something strange occurred. After I cleared out the data for that one account I found that the P&L cube depreciation numbers for a different depreciation account and one month were almost double the rule based number showing in the depreciation cube. When I performed trace cell it actually showed the cell result of double and the source data as different (source data matching to the rule based value in the depreciation cube). Making the story short I found that the depreciation cube views were showing the data as I expected, but the P&L cube was reading both the rule based depreciation data and the old non rule based values and adding it together. After removing all rules in the depreciation cube and clearing the forecast area and reintroducing the rules everything was working as expected. Has anyone come across a similar situation?

[Adam]

Admittedly, there’s a lot going on in your post.

What resonated with me the most was input data persisting after a rule is layered on top of it.

There is a tm1s.cfg setting that controls this exact behavior.

RulesOverwriteCellsOnLoad
https://www.ibm.com/docs/en/planning-an ... onload.htm

Assuming this setting is undefined the data persist. Good news is built into this setting is a cleanup on db startup.

I vividly remember discussing this topic with my IBM teacher when I was first learning TM1. He mentioned to me they were forced to add this setting to the product because an important customer lost a lot of data by accidentally applying a rule over input data. Yikes.

Hope this clears up most of your issue.

[JohnO]

There is a tm1s.cfg setting that controls this exact behavior.

RulesOverwriteCellsOnLoad
https://www.ibm.com/docs/en/planning-an ... onload.htm

Thanks, I wasn't aware of this one. I was sure that in previous experience a server restart had cleared old data and it now makes sense why that would have been the case.
RulesOverwriteCellsOnLoad = T it will be and as a general statement that should prevent most issues for us though the underlying data will persist until the server is restarted which we do once a week.

I wonder what they have done in Server 12 given with its memory management and logging and (Optional pay for it failover) it's not supposed to be a regular restart .

And after knowing about this setting I found this very old thread
viewtopic.php?t=7278

[Adam]

Glad I could help.

Calling back to the same onboarding to TM1, I was told we should never have to restart the DB except when the OS needs to restart.

Any reason you’re restarting the DB weekly?