Issue with feeders

[GeorgeB]

I'm currently having an issue with feeders that I can't explain - my only resolution so far is to turn SKIPCHECK off (behaviour as expected), however I believe this isn't ideal performance-wise.

I have a nodal intersection:


The rule for 'Butt Equivalents' is:
['Butt Equivalents']=N:
If ( !SCM Cask Size @= 'HHD', 0.5, 1 ) * ['Physical Casks'];

Since !SCM Cask Size @= 'Butt', 'Butt Equivalents' = 'Physical Casks' in this case.

The associated feeder is:
['Physical Casks']=>['Butt Equivalents'];

As we can see, 'Butt Equivalents' is not being fed, and this is confirmed when I 'Check Feeders'.

I have used CubeProcessFeeders() to no avail.

The behaviour that 'Butt Equivalents' is still showing despite zero suppression on both rows and columns, I think, is due to PersistentFeeders=T at some point - this is now PersistentFeeders=F, all associated feeder files have been deleted and the service has been restarted since, but this isn't my main issue.

My question to you guys is whether or not I'm being stupid here and missing something obvious, or if there is something going on behind the scenes I'm not aware of.

Thanks for any advice!

George

[Wim Gielis]

Hi George,

At first sight this should just work.

Did you do ‘Trace feeders’ starting in the white cell ?

Do you have a rule at C level to overwrite the top right cell to 0 ?

Do you have a Feeders section ?

BTW, I don’t understand this part: ‘Since !SCM Cask Size @= 'Butt', 'Butt Equivalents' = 'Physical Casks' in this case.’

[GeorgeB]

Thanks for the reply, Wim.

Did you do ‘Trace feeders’ starting in the white cell ?

I did trace feeders, the white cell is feeding several unrelated cells in a different cube but no cells within the same cube.

Do you have a rule at C level to overwrite the top right cell to 0 ?

I do not.

Do you have a Feeders section ?

I do, and the other feeders are working as expected.

BTW, I don’t understand this part: ‘Since !SCM Cask Size @= 'Butt', 'Butt Equivalents' = 'Physical Casks' in this case.’

This was only justification as to the result of the calculation under 'Butt Equivalents' - information only.

Thanks,
George

[declanr]

I think there used to sometimes be an issue if the time stamp on the blb? generated by the advanced rules editor came out of sync with the rux file. It’s been ages since I have seen it so it’s a long shot but I do think I saw it on site with you guys so might be worth a look.

[Steve Rowe]

Also agree this should be working.
Maybe do a rule trace on the zero consolidation and just check that there is no rule applying?

[GeorgeB]

Thanks for your replies, Declan and Steve.

by declanr » Thu Jun 21, 2018 7:22 pm

I think there used to sometimes be an issue if the time stamp on the blb? generated by the advanced rules editor came out of sync with the rux file. It’s been ages since I have seen it so it’s a long shot but I do think I saw it on site with you guys so might be worth a look.

The .blb and .rux files are in sync in this case - none of our developers are using the advanced rules editor as we we aware of the potential sync issues.

by Steve Rowe » Fri Jun 22, 2018 9:49 am

Also agree this should be working.
Maybe do a rule trace on the zero consolidation and just check that there is no rule applying?

The calculation on the zero consolidation is just "Consolidated: 0" - standard behaviour when no rule applies and the cells below it aren't fed.

[mattgoff]

Two suggestions:

• Stop server, delete .feeders files (if you use persistent feeders), restart server. Even if you're not using persistent feeders I've seen this resolve weird rule anomalies.
• Post or attach complete rule file, not an excerpt, which can help us rule out a problem elsewhere that's killing compile before it gets to the feeder.

Matt

[jcr55]

Two suggestions:

• Stop server, delete .feeders files (if you use persistent feeders), restart server. Even if you're not using persistent feeders I've seen this resolve weird rule anomalies.
• Post or attach complete rule file, not an excerpt, which can help us rule out a problem elsewhere that's killing compile before it gets to the feeder.

Matt

While the server is down, you could also delete the cube rule .blb file, just in case the .blb and .RUX files are not in sync.
Then when viewing the cube rule via Architect, you will ensure you are looking at the current rules.