TM1 Forum

Hi All - we had a significant issue following applying the latest HF3 (on advice from IBM it would fix a bug we had reported) where despite data in }ElementAttribute cubes being all OK the attributes were not "attached" to the dimension.
Symptom 1/ in subset editor attribute values display as "error" Symptom 2/ DBRW retrieval fails in Excel for any DBRW relying on an alias value for an affected dimension (this is the real one to watch out for!)

"local fix" is relatively easy being force dimensions to recompile by adding (and then removing) an element. Anyone else experienced any misbehavior or new regression bugs with HF3?

This error looks familiar but I got it a while ago (version 9.5 or prior) and what I recall I had to rebuilt the dimension. Did you check the }elementsAttributexxxx cube? I am assuming it would be showing you the same info as the screenshot.

rkaif wrote:This error looks familiar but I got it a while ago (version 9.5 or prior) and what I recall I had to rebuilt the dimension. Did you check the }elementsAttributexxxx cube? I am assuming it would be showing you the same info as the screenshot.

No that's the thing, the element attributes cubes were 100% fine. Moreover the alias display in the subset editor was also working as was returning an alias from SUBNM formula. I think this says something interesting as to how TM1 handles attaching attributes behind the scenes as although the alias values were in the cube and could be seen even in subset editor they were clearly not assigned correctly as aliases by virtue of formulas using the aliases failing (and the error display in subset editor.)

I have seen this before in 9.1.3 and 9.5.1 but only intermittently. I just hope this was a once off and not a regression bug with HF3.