extremely long execution while cube data view

[former_member245939]

Hi experts!

My issue:

I have infocube zEsncub. It contains only 100.000 records and 10 requests in admin panel.

while I am execute infocube data view in RSA1, my query to cube hangs in DB and doesn't give any reaction.

Selection Creterias:

zYear = 2015

zMonth = 09

zBus_area = 0001

BUT! If I delete char "zMonth" from Selection Criterias and execute, then query works fine.

All these chars are in one dimension of cube.

I can't undestand such behaviour. Please help me solve this problem.

Indexes and statistics in cube are green.

[Former Member · Accepted solution]

Hi Sergey,

Can you enable the 'Use DB aggregation"' and  then view the data ?

Regards,

Amit

[former_member245939 · Accepted solution]

Hi Amit!

Yes, i have checked 'Use DB aggregation' and have got the data. Thanks for helpful advice!

But normally I don't use this option while data view. And nevertheless I get data without problems (using other criterias) ...

But what is the reason ? In char "zMonth", or maybe in indexes of dimension table?

[Former Member · Accepted solution]

Hi Sergey,

Good to know you were able to get the values.

Now coming to the issue, it might be due to the fact that there are large number of records for 'zMonth'

It's better always to use 'DB aggregation while viewing data'

Regards,

Amit

[former_member245939 · Accepted solution]

No, Amit.

1. 'zMonth' contains SMALL number of records (12 months)

2. in both cases I want to get only 20 records, so theoretically this option doesn't matter ...

[KodandaPani_KV]

Hi,

can you create the aggregates on the cube for the zMonth.

or

create the indexes at cube performance tab.

[former_member183519]

Hello ,

Can u try to create secondary index based on zyear zmonth zbus_area ?

Go to RSA1> Infocube > Performance > Create Index

let us know performance after that.

Regards,

Hitesh Gavande

[RafkeMagic]

are your InfoCube statistics up to date?

(RSA1 > InfoCube > Manage > Performance > Check Statistics ==> this should ALWAYS be green)

[former_member245939]

Hi Raf,

yes, statistics and indexes are Green

[Former Member]

Hi,

Instead of having all the characteristics in one dimension, you can split the characteristics and define the dimensions, "Multi dimensional modelling".

Regards,

Priya

[former_member245939]

Hi Priya,

I agree with you, Multi dimensional modelling is right way of Cube design

But it's Prod system and this cube is very old. There were no such issues earlier

These chars have small SID table and cube is small (100.000 recs)

Looks like DB issue...

[Former Member]

Hi Sergey,

You can go with indexing the cube, you can delete the index and recreate it.

Regards

Vivek

[former_member245939]

Hi Viveklal,

I have tried ro do these actions, unfortunatelly no effect...