Martin Schoombee (4/10/2015)

---

VRT (4/10/2015)

---

Thank you for your reply Martin.

I am already using this document as reference and worked on this ROLAP cube. I was really happy when I first see the data coming up in 2 minutes using Excel Pivot. But later it started giving trouble.

We already used PowerQuery and PowerPivot tools and those are really a nice tool as per their standards. We do use the PowerBI tools for our dashboarding needs by writing custom SQLs.

But we have our data warehouse tables which are big in width and depth.

150 columns and more than 100 million records each. So I went for ROLAP concept since In-Memory concept wont work out in this case.

Understood, but why not use MOLAP (or at least HOLAP). If you have a close to real-time requirement, you could consider switching in partitions for data that recently changed.

In my opinion, any tool would struggle to aggregate hundreds of millions of rows over very large dimensions. You probably either need to reduce the scope (i.e. multiple smaller cubes), or find somebody with a heavy check book and spend a few million.

One last suggestion. Depending the version, have you considered using Oracle's equivalent of columnstore indexes for some of these large tables? It would definitely speed up any relational query significantly, if you could live with the nuances.