Search found 524 matches

by Duncan P
Wed Nov 15, 2017 5:36 pm
Forum: IBM TM1, Planning Analytics, PAx and PAW
Topic: Negligible value in consolidation
Replies: 2
Views: 246

Re: Negligible value in consolidation

If you are interested in why it happens then have a read of this.

The effect is magnified in consolidated numbers and the level of error increases the more leaf numbers are involved.
by Duncan P
Sun Jul 16, 2017 8:57 pm
Forum: IBM TM1, Planning Analytics, PAx and PAW
Topic: negative value divided by zero Issue
Replies: 3
Views: 476

Re: negative value divided by zero Issue

Out of interest what happens when the denominator is the extremely small number used in place of null in some instances when we have UNDEFVALS? I would hope that it would realise and treat it as zero.

It's been a while. Is UNDEFVALS still a thing out there?
by Duncan P
Sat May 20, 2017 8:46 pm
Forum: Useful code, tips and tricks
Topic: Double counting elements in a dimension
Replies: 5
Views: 1194

Re: Double counting elements in a dimension

Wim, you've made my day. There's someone else out there that puts the comma at the beginning of each parameter line instead of the end.
by Duncan P
Tue Apr 11, 2017 6:04 am
Forum: IBM TM1, Planning Analytics, PAx and PAW
Topic: TM1 10.2.2 FP6 IF1 crashing without log entries
Replies: 10
Views: 1101

Re: TM1 10.2.2 FP6 IF1 crashing without log entries

Given that you are running on Windows have you looked in the Windows Application event logs to see if the process dying is leaving any trace in there?

"Programs -> Window Administrative Tools -> Event Viewer" or something similar.
by Duncan P
Tue Apr 11, 2017 5:58 am
Forum: IBM TM1, Planning Analytics, PAx and PAW
Topic: How to solve SystemOutOfMemory errors when running reports?
Replies: 7
Views: 1243

Re: How to solve SystemOutOfMemory errors when running reports?

This APAR might be relevant depending on the version of BI you are using.
by Duncan P
Mon Apr 03, 2017 2:12 pm
Forum: IBM TM1, Planning Analytics, PAx and PAW
Topic: Export Cube data in CSV file, with all Months as columns
Replies: 22
Views: 2098

Re: Export Cube data in CSV file, with all Months as columns

Let feeders be your friends. http://www.tm1forum.com/viewtopic.php?f=3&t=7439

Feeders will not suffer from the cancelling out of normal accumulation.
by Duncan P
Fri Mar 17, 2017 4:42 pm
Forum: IBM TM1, Planning Analytics, PAx and PAW
Topic: Number Consolidation based on security settings
Replies: 7
Views: 781

Re: Number Consolidation based on security settings

Even in Cognos Planning the access tables did not affect the way in which the consolidation was done. If there data was there it was aggregated, even if it was invisible to the user. We had arguments internally about the problem Steve highlights but there was never any question of doing the aggregat...
by Duncan P
Wed Mar 15, 2017 9:09 am
Forum: IBM TM1, Planning Analytics, PAx and PAW
Topic: fact table
Replies: 3
Views: 500

Re: fact table

So if you are talking about a star-schema structured fact table in a dimensionally modeled relational warehouse (Kimball style) then TM1 does most of this intrinsically. A Dimension in TM1 corresponds to the dimension table in the star schema. The key difference is that whereas the individual levels...
by Duncan P
Tue Jan 03, 2017 12:31 pm
Forum: IBM TM1, Planning Analytics, PAx and PAW
Topic: TM1 replacement for Cognos Controller
Replies: 1
Views: 320

Re: TM1 replacement for Cognos Controller

It depends whether you are asking "Could you implement a consolidation package in TM1?" or "Could I just migrate from Controller to PA?" The answer to the first is yes - various others here can point you at particular offerings. It is even possible to reproduce some exact Controller behaviour in det...
by Duncan P
Fri Dec 09, 2016 8:54 pm
Forum: IBM TM1, Planning Analytics, PAx and PAW
Topic: Server crash after data entry
Replies: 8
Views: 854

Re: Server crash after data entry

No idea if this is still the case but at one time if you had a circular reference that was enabled conditional on some data entry, and if the circularity involved multiple cubes then the server didn't detect it and blew up with stack full.

As I said I have no idea if this is still the case.
by Duncan P
Wed Sep 07, 2016 11:07 am
Forum: IBM TM1, Planning Analytics, PAx and PAW
Topic: Rule on consolidated level and slicer set
Replies: 4
Views: 555

Re: Rule on consolidated level and slicer set

I remember an issue where the Cognos BI infrastructure could not cope with a rule-defined consolidated value for which none of the underlying leaf level values were, or evaluated to non-zero, regardless of whether it was fed or not. Could this be your problem? I had to recommend a workaround where a...
by Duncan P
Fri Sep 02, 2016 7:54 am
Forum: IBM TM1, Planning Analytics, PAx and PAW
Topic: SubsetElementGetIndex TI function
Replies: 9
Views: 986

Re: SubsetElementGetIndex TI function

The key thing is that by specifying a start index they have given a way of iterating over all the occurrences of an element in a subset.
by Duncan P
Sat Aug 20, 2016 6:11 pm
Forum: IBM TM1, Planning Analytics, PAx and PAW
Topic: ConsolidatedAvg with several dimensions
Replies: 2
Views: 384

Re: ConsolidatedAvg with several dimensions

ConsolidatedAvg gives you the average of all leaf cells below the one you have chosen with the coordinates. It appears that you want the average of the sums across client. You should be able to get this simply by multiplying each result by the number of leaf clients feeding into it. You can get that...
by Duncan P
Thu Feb 11, 2016 8:08 am
Forum: IBM TM1, Planning Analytics, PAx and PAW
Topic: TI processing locking file
Replies: 7
Views: 651

Re: TI processing locking file

To be sure you need ProcMon. This will tell you exactly which process is locking the file.
by Duncan P
Mon Feb 01, 2016 8:24 am
Forum: IBM TM1, Planning Analytics, PAx and PAW
Topic: Huge amount of memory used for input data
Replies: 5
Views: 851

Re: Huge amount of memory used for input data

qml wrote:To me it looks like in your case 'Memory Used for Input Data' and 'Memory Used for Feeders' are both wrong
this is highly likely particularly given that
its_me wrote:Memory Used for Feeders 16777216
is an exact power of 2.
by Duncan P
Thu Jan 07, 2016 8:25 pm
Forum: IBM TM1, Planning Analytics, PAx and PAW
Topic: TM1 Rule file unresponsive
Replies: 12
Views: 1443

Re: TM1 Rule file unresponsive

It may be worth checking whether any of the feeders are targeting consolidated cells. If so then you might benefit from reading this.
by Duncan P
Sun Dec 13, 2015 11:01 am
Forum: IBM TM1, Planning Analytics, PAx and PAW
Topic: TM1 10.3 - IBM Planning Analytics
Replies: 40
Views: 12025

Re: TM1 10.3 - IBM Planning Analytics

The MTQ=All parameter is interesting. Do we know if that means physical cores or hyper-virtual cores? The documentation here did not specify.
by Duncan P
Sun Nov 22, 2015 1:15 pm
Forum: IBM TM1, Planning Analytics, PAx and PAW
Topic: tm1s-log.properties for Process Parameters
Replies: 7
Views: 1545

Re: tm1s-log.properties for Process Parameters

It's a good one for an RFE. But until that comes along ... We used to have our own additional logging and a convention that all processes would log their parameters as the first thing they did when they executed. I was working with a very organised chap at the time who set up all the infrastructure....
by Duncan P
Sat Nov 21, 2015 9:28 pm
Forum: IBM TM1, Planning Analytics, PAx and PAW
Topic: tm1s-log.properties for Process Parameters
Replies: 7
Views: 1545

Re: tm1s-log.properties for Process Parameters

It depends how your processes are being called, but if you are interested in parameters for processes called from within other processes they will appear as function arguments to the ExecuteProcess function in the calling process.
by Duncan P
Mon Nov 09, 2015 12:29 pm
Forum: IBM TM1, Planning Analytics, PAx and PAW
Topic: How to prevent a commit / force a rollback
Replies: 16
Views: 2622

Re: How to prevent a commit / force a rollback

If you wanted N multi-threaded processes to commit or rollback in coordination (if such functionality were possible) you would have to have them counting success/failure flags and only commit when all N (including themselves) have been counted. This is known in the trade as a " two phase commit prot...