1

We ha a Fact table with 500 milions of rows (data from 10 years - 60GB). I've partitioned this table on database level and SSAS Tabular Cube level as well - one month per partition to improve the overall browsing performance and SSAS Cube processing time. But last time I've heard that configuring more than 2 partitions for SSAS Tabular cube is not good. There should be max 2 partitions per tabular cube. Have you heard about something like that ? Is that true ?

1

In SSAS 2014 Tabular partitions inside a table process serially. In SSAS 2016 Tabular partitions inside a table can process in parallel. So in your version partitioning won't speed up processing if you process the whole table.

However partitioning is great for use in incremental processing. If you partition by year and only 2017 rows changed then you can just process the 2017 partition. Also to archive off data older than 10 years old you can drop the oldest partition.

I don't know where the 2 partition rule you mentioned came from but don't believe it makes sense. If you have a reference for that assertion with some context then feel free to share.

  • Thanks. Unfortunately I don't posess any exact link to such article. I've just get that information from one of the BI consultant last time. – Tomasz Wieczorkowski Jun 6 '17 at 14:32
  • @TomaszWieczorkowski if I have answered your question please consider accepting it by clicking the check-mark. – GregGalloway Jun 7 '17 at 10:44
  • Oh yes, here you go :) – Tomasz Wieczorkowski Jun 7 '17 at 11:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.