Forum Discussion
Refresh error / Exceed maximum page size of the storage object (65528 bytes)
After extensive search, we came to the conclusion, that the dataset had too many columns. The number of rows wasn't an issue—it was all about the columns. This, combined with the architecture of the dataset, caused the problem.
We are loading multiple tables with this structure:
Row Nr. Item 01/31/2021 ..... 12/31/2045 1 Europe 0,05 ..... 2,5 2 Asia 2,5 ..... 2,6 .... .... .... ..... .... 10.000 NA 5,5 ..... 5,4
The dataset architecture used queries for each of these tables. In the second step, we merged all these queries into a master table and repivoted all the columns from the matrix format into one table with only four columns:Row Nr. Item Date Value 1 Europe 01/31/2021 0,05 .... .... .... .... 1 Europe 12/31/2045 2,5 .... .... .... .... When reloading the OLAP Cube, the error message popped up.
How did we fix the issue?
We repivoted each table before merging them into the master table. By doing this, we prevented the OLAP Cube from loading an excessive number of columns.An alternative, maybe easier approach is - if the multiple tables are being merged into one table and repivoted (like in the example above) - deactivating the single Tables/queries from being loaded into the dataset also helps. The data is still being considered in the master table, but the single Tables/queries are spared from entering the OLAP Cube.
Hope that helps someone.
Super UserI assume you want to stick with MDX and do not want to consider running DAX queries instead?
Unfortunately, setting an entire new model (DAX) isnt an option. I dont get why multiple queries with a lot of columns and rows doesnt cause an issue, but deleting them and implemeneting an single query with less columns and rows does...
- lbendlin
When you connect to an OLAP cube you connect to all of it. Performance will depend on which columns you choose and how they are related in the multi dimensional data model. It is very easy to bring a cube to its knees by choosing only a few but unrelated columns.
Thanks for your feedback. I checked but couldnt find any unrelated columns in the model. Is it possible to load only selective data into the OLAP cube? If not, is there any other approach i could try?