Ultimate Long TM1 quiz

[EP_explorer]

Anybody else? I think it's one of the hardest questions in this quiz

@MVPs, @Admins, can you share your thoughts also?

As we have to feed ‘a’ all the time (4 is unchanged) so
add new cube with input value, or element 'c' - input all values in it (1 for example)
and
[‘c’] => [‘a’];

[lotsaram]

Anybody else? I think it's one of the hardest questions in this quiz

@MVPs, @Admins, can you share your thoughts also?

Code: Select all

[‘a’] = [‘b’] + 4;

Well this specific case is highly, highly theoretical as I can't imagine any practical use for such a rule in real life. Seem as it is an open rule applying at any level I would argue it doesn't need a feeder at all and it would be better to simply define "a" as a consolidation with "b" as its child. You could then formulate the rule as

Code: Select all

[‘a’] = C: [‘b’] + 4;

... which doesn't require feeding. Yes there might be edge cases where "b" is empty and so "a" might disappear under null suppression. However, in a real world scenario there would be multiple other dimensions at play in the cube, and the higher you aggregate on any of them then the more likely that there is some data somewhere on a slice containing "b". Hence for practical purposes feederless C rule is the way to go. And for edge cases where "b" is empty then turn null suppression off and "a" will show the correct result. And showing the correct result is the whole point (only point) of feeders in the first place.

[Adam]

I have collected a huge amount of TM1 questions while interviewing people. I want to share them with the community, and I want to make this in a form of a quiz. After all this ends, everybody will be able to use all questions for interviewing or for interview preparation.

I don't know about you guys, but I am treating this quiz like a legitimate interview.

Write a feeder for this rule:

Code: Select all

[‘a’] = [‘b’] + 4;

This is the interview question. While I agree that removing SKIPCHECK or making this a FEEDERless C are all valid alternates, they don't actually answer the interview question "write a feeder". Rather, they answer a different question all together (optimization).

Therefore I'm sticking with my answer.

[Elessar]

Correct answer:

Well, there is no correct answer here in technical point of view. Rule itself is a mistake. And if this is your rule, you lose. If you inherited this rule from somebody else, you can punch him.

if you see such a rule as a requirement (and not yet written it in such way), you need to

1. Stop working with analyst who has given you such a requirement
2. Send another analyst to analyze what does this "4" mean, is it always "4" and what the hell is going on here

And obviously you cannot just feed from ['b'].

I have faced this once: when a "rather experienced" Analyst has taken "= C2 * AA12 + 13000" (where 13000 is obviously adjustment which "rather experienced" Excel users like to use in such a way) Excel formula as is and put it to the requirements, and "rather experienced" TM1 developer has taken requirement as is and put it to rule.
And definitely, in this case, we need another element for this adjustment and feed from it (or just include it to consolidation and make it feederless).

Winner of this week is Adam and he takes the lead!

Question #20:

You have a cube with detailed costs data and a reporting cube where 1st cube's consolidated accounts are leaves. Write a rule and a feeder to pass data from 1st cube to 2nd cube

Image 54.png (98.31 KiB) Viewed 10993 times

[EP_explorer]

question20.JPG (71.6 KiB) Viewed 10936 times

[MarenC]

Hi,

As an alternative to ep_Explorer solution, I might be tempted to add an attribute to the Account dimension called, let us say, 'Feed Parent' and use that instead ELPAR, as an account could have more than 1 parent. And populate the attribute via TI as part of dim update process. Also ensure reprocess feeders is inserted where applicable!

Maren

[jwilliamson_uci]

Hi,

I could be mistaken, but wouldn't you also have to reference the context dimensions? Without knowing the exact order of the dimensions, shouldn't the resulting rule be something like:

SKIPCHECK;
[] = N: DB ('600_Sum_Details', !600_Sum, 'Total Organization', 'Total Line Item', !months, 'Amount');

I'm guessing that there is a measure dimension offscreen, and I've used "Amount" as my measure. The feeder remains as EP wrote it unless you want to use Maren's suggestion of an attribute.

Jeff

[Elessar]

sorry for being late: was on vacation

Correct answer:
Will be this Friday, I think.

EP_explorer's rule works, but not in a way I would use: ELPAR is rather dangerous function here, because new consolidations can be created before the consolidations on screenshot, and we will need ELPAR(X, X, 2). Solution with attributes is better and MarenC could have earned this point if he wrote a rule. Thus, EP_Explorer is winning this round, and let's create a new question based on previous:

Question #21:

Write rule and feeder for the same task, but using attributes.

P.S. and yes, the idea of this question is to use correct approach. Dimensions in context, their order are not so important.

[Elessar]

Anybody? I will need to postpone next question until this question is answered.

[EP_explorer]

The MarenC's idea looks like

04.jpg (171.14 KiB) Viewed 9592 times

[EP_explorer]

In defence of idea of Elpar I could say:
The dimension of sum elements in conditions of the tasks seems is based on the dimension of sum and details elements
And if you add new sum element in sum and details dimension you have to after it change dimension of sum elements and elpar continue works if the structure of dimensions will not change.

If structure of dimensions will change - new levels appear - yes, using attribute is more resilient

[MarenC]

Hi,

I somehow missed Question 21! I was just waiting for you to provide the correct answer. But I had assumed ep_explorer took the points.

Maren

[Elessar]

Correct answer:
Is described here: viewtopic.php?p=83624#p83624
Also notice the correct mapping for feeder: it should be for each detailed element, not for consolidations themselves.

Thus, EP_Explorer is winning this round again!

Question #22:

In which cases you DO NEED to refresh security?

• Security is based on }ElementSecurity cube rule. Rule is changed.
• Security is based on }ElementSecurity cube rule. New element is added to the dimension
• }ElementSecurity cube does not have any rules. Particular cell is changed manually
• }ElementSecurity cube does not have any rules. Security is changed using «ElementSecurityPut» TI function.
• Security is based on }CellSecurity cube rule. Rule is changed.
• }Cellsecurity cube does not have any rules. Particular cell is changed manually

[ardi]

• Security is based on }ElementSecurity cube rule. Rule is changed. NO
• Security is based on }ElementSecurity cube rule. New element is added to the dimension YES
• }ElementSecurity cube does not have any rules. Particular cell is changed manually NO
• }ElementSecurity cube does not have any rules. Security is changed using «ElementSecurityPut» TI function. NO
• Security is based on }CellSecurity cube rule. Rule is changed. NO
• }Cellsecurity cube does not have any rules. Particular cell is changed manually NO

[Elessar]

Spoiler: one bullet is incorrect in Ardi's answer.

[michalb]

Security is based on }ElementSecurity cube rule. Rule is changed. YES
Security is based on }ElementSecurity cube rule. New element is added to the dimension YES
}ElementSecurity cube does not have any rules. Particular cell is changed manually NO
}ElementSecurity cube does not have any rules. Security is changed using «ElementSecurityPut» TI function. NO
Security is based on }CellSecurity cube rule. Rule is changed. NO
}Cellsecurity cube does not have any rules. Particular cell is changed manually NO

[Elessar]

Correct answer:
Is described here: viewtopic.php?p=83638#p83638

Leader of this round is michalb. Congrats and welcome!

Question #23:

GroupsCreationLimit - where is this parameter set? What are recommendations for optimal parameter value?

[ardi]

Correct answer:
Is described here: viewtopic.php?p=83638#p83638

Leader of this round is michalb. Congrats and welcome!

Question #23:

GroupsCreationLimit - where is this parameter set? What are recommendations for optimal parameter value?

GroupsCreationLimit Is a tm1s.cfg parameter. TM1 uses this param value during restart to allocate memory, so a very large number can lead to out of memory errors.
So i would not set this value more than 200 but it all depends on the memory you have on the box

[Elessar]

Correct answer:
The parameter is depreciated in 10.2.2 FP3. I've added this question because too many oldies are asking me "why is GroupsCreationLimit not set in your tm1s.cfg". And Ardi got caught! If anybody wrote about depreciation also, he would take this point.

But nobody did, so winner of this round is ardi!
For older versions, everything is correct. My optimal value was about 30 for productions server.


Question #24:
Two different picklists are set in "picklist" attribute and }picklist cube. Which one will work?

[MarenC]

Hi,

Two different picklists are set in "picklist" attribute and }picklist cube. Which one will work?

If the Picklist cube value is not blank then it will take the picklist cube value. If the picklist cube value is blank the picklist attribute value is not blank then it will take the attribute value.

Maren