TM1 Forum

Here's a feeder conundrum. Warning: it's a bit tedious to read through; my apologies if it's too verbose to make the point.

I thought that when you unload a cube, its feeders are destroyed. Suppose then, you have two cubes joined by rule/feeder. Let's call the one with the feeder Cube1 and the one with the rule Cube2.

If you unloaded both cubes, then reloaded just Cube2, wouldn't you expect the inter-cube number not to consolidate? In fact what happens is that at first load, it doesn't consolidate, but when you press F9 it does! Suggesting that somehow the feeders are re-set from the (still unloaded) Cube1 just by pressing F9.

Even weirder, if you unload Cube2 again and re-load it, this time F9 does not consolidate the number. Can anyone explain the mechanics of this?
I thought I would need to space-bar and re-save the rule in Cube1 for the inter-cube feeder to be re-generated (we are on v9.0). Is it related to cacheing and dependencies? I haven't been able to figure it out yet.

Here's the scenario:
a. Create two cubes with an intercube rule/feeder. Cube1 has the feeder; Cube2 has the rule.
b. Unload both cubes (right-click, Unload cube)
c. Reload Cube2. Result: No conoslidation calculated (expected). Press F9 - consolidations are calculated (unexpected).
f. Unload Cube2 again.
g. Reload Cube2. Result: No conoslidation calculated (expected). Press F9 - consolidations are not calculated (???)

Regards, Ian

I know back in the day Applix made efforts to make sure rules were compiled within the cube so that is the RUX fie was destroyed by accident the rules could be recovered. It could be something to do with that? (Stretching) The only other question is do you have persistent feeders enabled??

iansdigby wrote:(we are on v9.0)

Wow. What is that now, almost 15 years old? I can't remember what a did a few years ago, much less 15. So much has changed since that was first released.

tomok wrote:

iansdigby wrote:(we are on v9.0)

Wow. What is that now, almost 15 years old? I can't remember what a did a few years ago, much less 15.

I missed that bit. You're right wow. Persistent feeders??? I don't think they were even a twinkle in Dave Corbett's eyes back then.

jim wood wrote:I know back in the day Applix made efforts to make sure rules were compiled within the cube so that is the RUX fie was destroyed by accident the rules could be recovered. It could be something to do with that? (Stretching) The only other question is do you have persistent feeders enabled??

Thanks for the comments, all. It is (perversely) comforting that such knowledgeable persons as yourselves also don't know. Yes, v.9 is very old, so I'm having to go back to basics. Jim, your interesting comment has the ring of truth about it.

I did some more experimentation and I (think) discovered the following:

If you manually unload a 'source' cube (i.e. one that feeds another) from memory in v9.0, TM1 generates the feeders in the dependent cube as it unloads. Presumably by virtue of the dependency table it maintains. I tested this many times and it seems to hold true.

By 'source' and 'dependent' I mean the opposite of what the help files counter-intuitively imply.

Caveat: this behaviour may not apply to other versions, e.g.if you manually unload a 'source' cube, TM1 may appear to do it but may not actually, as implied by lotsaram here:
http://www.tm1forum.com/viewtopic.php?f ... ers#p34800

Ian