[Mondrian] <join>,<Query> and <view> for snowflake schemas
swucareer99 at gmail.com
Mon Mar 21 16:30:09 EDT 2016
Thank you so much for spending time on the testing and replying
On Mon, Mar 21, 2016 at 4:21 AM, Isaias Sanchez <isaias.sanchez.l at gmail.com>
> I've tested all and all three options perform more or less the same. The
> difference can be on memory available for database cache or the one
> available for mondrian cache, also if you want to use some database tricks
> for the join is preferable to use <View> or <Query> (these two are the same
> in performance in my opinion).
> Talking about cache probably you have some dimension table with 10 columns
> and some unused or expired rows. If you use <Join> mondrian will take all
> that data even if it won't be used in the cube. With a View or a Query you
> can filter that.
> As a test I recommend you create your Query and check saiku or database
> logs to check the Query created by <Join> and compare their performance
> against each other, there you will have the better option.
> Isaias S.
> Selina Tech wrote on 19/03/16 00:32:
> Dear All:
> When I study the Mondrian schema. I think we have three different ways
> for the Dimension of snowflake schema.
> Way 1: Use <Join> as the example at
> Way 2: Use <Query> to define a ‘table’, and use this ‘table’ for Dimension
> Way 3. Use <View> to define a view ‘table’ by <SQL> and use this ‘table’
> for Dimension table
> Is any one of them wrong? if they are right, how could I choose one
> of them? How is the different on performance? when we use <view>, is this
> <view> will generated in memory, disk or database? what happen if the
> result of view is huge on size?
> Your any help is highly appreciated.
> Mondrian mailing listMondrian at pentaho.orghttp://lists.pentaho.org/mailman/listinfo/mondrian
> Mondrian mailing list
> Mondrian at pentaho.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Mondrian