Excel Pivot Tables on Impala via MDX... Intrigued?? Check out our video teaser

That is very impressive! Aside from the obvious questions about when and how this might be available, can you shed any light on what sorts of MDX capabilities you've been able to translate into HiveQL, or perhaps more importantly what is not supported? If you'd like some help beta testing this at some point, please feel free to contact me directly as we have previous experience with Simba (connecting to SAP HANA) and more recently have been spending a lot of time with Impala.

Cheers,
John Tyler

Hi John, thanks for your kind words and for your offer to test! The
software is in early development and we’re not quite at the stage where we
can talk about availability. We really appreciate your feedback and any
suggestions you have for us. Suggestions in terms of the types of data sets
you want to analyze, the MDX functionality you want to use, and Excel
features you want to use would be very helpful.



We cover the breadth of functionality in MDX. This includes both MDX query
support and MDX calculation support for those calculated members and
measures inside the cube. We adapted Simba’s MDX engine, which translates
MDX->SQL, to do the MDX->HiveQL translation. HiveQL supports almost all SQL
constructs, so this adaptation was not too difficult. There are no
meaningful restrictions in what you can do with MDX to be aware of at this
time.

It seems we will be having multidimensional structure underlying the Impala
table structures .So Wont it be any performance issue , since we are going
to convert the MDx query into the HiveQL and sends the HiveQL to Impala,
collects the results and returns them to the end user.

So, At peak usage time wont the system with more users will tend to suffer
more performance problems if I wasn’t wrong on this. Hope to see and use
this great tool in near future.
Thanks,
Anil Maharjan

Hi Anil,

Thanks for the note.

We do not expect any direct performance issues with the generated HiveQL
due to Impala's performance.

With regards to peak time and high concurrent user load, that will be
Impala's responsibility which I understand they're working on.

George

This is very interesting and certainly something that would be valuable in
the market. How and where are the "cube" definitions that the MDX
references mapped to the underlying Impala table structures?

Hi Nenshad, Simba has a tool targeted at developers for building cube
definitions, which we call a schema. Conceptually, this tool is similar to
the one for building Analysis Services cubes. The cube definition written
in Simba’s tool generates some XML that maps MDX metadata constructs to the
Impala table structures. In this demonstration we store the XML in a flat
file; the final location of the schema for a full version of the product is
yet to be determined.

Hi Simba Team,
     first off this is excellent work ! it makes total (business and
technical) sense to have excel pivots connect via the cubes to the bigdata
impala tables.
given that it has been about a year and a half, is it live now by any
chance? if not i'd be happy to beta test this out and provide feeback.
please let me know


Thanks
Kaushal Jha


To unsubscribe from this group and stop receiving emails from it, send an email to impala-user+unsubscribe@cloudera.org.