TM1 Forum

Hi there,

I have a problem with the RestAPI

If I export a view via Ascii Export I get 34960 lines back.

If I query the view via the RestAPI, I get 34013 lines back.

I call the RestAPI via a Post:
https://MyServer.de:8001/api/v1/Cubes(' ... xpand=Axes($expand=Hierarchies($select=Name),Tuples($expand=Members($select=Name))),Cells

Does anyone have an idea what I am doing wrong?

I work with PA 2.06.

Thanks a lot!
orlando

Hello,

What kind of cells are contained in the view ? Level 0 / consolidated / string. Rules-calculated or not ? Fed or not ?

Can you make the view smaller to identify the gaps and find a pattern ?

Hi Wim,

thanks for the reply.
The view contains only Level 0, no strings and the cube has no rules.
I will try to make it smaller and investigate.
This will take a while, a lot of meetings and vacation ahead

Best regards,
orlando

You have a rather good analytics engine available. I suggest you create a new cube with a measure
Ascii
Rest
Difference = Ascii - Rest
and import the two feeds. You should then be able to identify the differences.

Or open up both exports in Excel, one below the other. Make sure the same columns are used in the same order. Then multiply 1 set of values with -1 and create a pivot table. The differences will remain.

Did you re-order the cube dimensions ?
There is a bug in PA (can't recall which release) with Rest API that will cause such issues if the cube has a re-ordered dimensions. Even PAx is facing the same problem.

kangkc wrote: ↑Tue May 21, 2019 1:57 am Did you re-order the cube dimensions ?
There is a bug in PA (can't recall which release) with Rest API that will cause such issues if the cube has a re-ordered dimensions. Even PAx is facing the same problem.

Hi,

i used an VLookUp in Excel to identify the missing rows.
The data doesn't look unusual and yes I re-ordered the dimensions.
We are still in development, so destroy the cube and reorder the dimensions could be an option - maybe

I'm in Ehingen at the "Planning Analytics TM1 Anwendertag" next week - i will ask there and come back with a solution after my vacation.

Thank you all for the ideas!

Best regards
orlando

orlando wrote: ↑Thu May 23, 2019 12:21 pm Hi,

i used an VLookUp in Excel to identify the missing rows.
The data doesn't look unusual and yes I re-ordered the dimensions.
We are still in development, so destroy the cube and reorder the dimensions could be an option - maybe

I'm in Ehingen at the "Planning Analytics TM1 Anwendertag" next week - i will ask there and come back with a solution after my vacation.

Thank you all for the ideas!

Best regards
orlando

You don't need to destroy the cube to re-order.
Re-ordering the cube back to the original dimension order can be on the fly without re-creating the cube.

kangkc wrote: ↑Fri May 24, 2019 12:56 am
You don't need to destroy the cube to re-order.
Re-ordering the cube back to the original dimension order can be on the fly without re-creating the cube.

Thanks for the tip
I wouldn't have thought that TM1 would recognize this.
Strange bug.

I will talk to them on tuesday - and come back with the answer.

I will also be at the Anwendertag next week. If you are interested in re-ordering (aka "cube optimization") and what it does I recommend you read these posts.
what is cube optimization?
how to find optimal dimension order

lotsaram wrote: ↑Fri May 24, 2019 7:40 am I will also be at the Anwendertag next week. If you are interested in re-ordering (aka "cube optimization") and what it does I recommend you read these posts.
what is cube optimization?
how to find optimal dimension order

hi,
thanks for the links.

See you in Ehningen