Great article <https://analyticsengineers.club/wha...
# data-newst
taylor
08/18/2021, 4:22 PMGreat article https://analyticsengineers.club/whats-an-olap-cube/
a
aaronsteers
08/18/2021, 7:25 PMI am a huge fan of OLAP cubes for the performance, semantic, and usability benefits they offer. I have heavily used SSAS and I wanted an open source version like Mondrian to take off. I'm curious if there are other open source options people know of.
aaronsteers
08/18/2021, 7:26 PMInteresting that Clickhouse is in this list. I didn't think of them in the "Cube" category but admittedly I don't know much about them. In Comparison of OLAP servers - Wikipedia, I see they mostly support native SQL with some extensions, but no MDX or XML-A.
t
taylor
08/18/2021, 7:29 PMnow I feel like I don't understand "OLAP" - why wouldn't Snowflake/BigQuery/Redshift be on here
a
aaronsteers
08/18/2021, 7:35 PMOLAP doesn't necessarily mean "OLAP cube". All cubes are OLAP but not all OLAPs are cubes? 🙂
aaronsteers
08/18/2021, 8:33 PMThe OLAP Cubes page has a Comparison of OLAP Servers link but not a "Comparison of OLAP Cubes" page.
t
taylor
08/18/2021, 8:34 PMYeah, I guess I'd just been using "OLAP" to describe any data warehouse. "OLAP Cube" for me has always been a specific table implementation, but I didn't think it'd be limited to a specific set of databases/warehouses
a
aaronsteers
08/18/2021, 8:36 PMSnowflake/BigQuery/Redshift are databases but not analytics engines per se. The MOLAP/ROLAP/HOLAP distinction on the page linked above aludes to where aggregations are created and cached. In general, to be a "cube", the engine needs to understand that Days roll up to Month, roll up to Quarters and Years - and then the engine can pre-aggregate data at any of those levels.
aaronsteers
08/18/2021, 8:36 PMThe OLAP Cube should be able to pre-aggregate data at any level, because it natively understands there's a natural hierarchy.
t
taylor
08/18/2021, 8:36 PMah ok - interesting distinction
a
aaronsteers
08/18/2021, 8:36 PM👍
t
taylor
08/18/2021, 8:37 PMthanks for detailing 🙂
a
aaronsteers
08/18/2021, 8:37 PMI want to ask Claire to do her next article on a comparison of OLAP Cube providers 😄
aaronsteers
08/18/2021, 8:37 PMAnd now I want to see if Clickhouse, for instance, meets that criteria - I actually don't think it's a "Cube" although it certainly seems tuned for OLAP workloads
4 Views