lens-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "amareshwarisr ." <amareshw...@gmail.com>
Subject Re: Purpose of creating Storage
Date Wed, 03 Jun 2015 05:14:03 GMT
On Tue, Jun 2, 2015 at 10:38 PM, Srinivasan Hariharan <
srinivasan.hariharan@outlook.com> wrote:

> Hi ,
>
> Is it mandatory to create storage specifficaly for each set of cubes,facts
> and dimesnions?.
>

Creating storage is required for two reasons.


   1. To associate physical tables to logical Fact and Dimension tables.
   2. To associate which drivers  can work which storages. Driver
   configuration would say, supported storages are "s1,s2", if s1 and s2 are
   storage names.



> For example say I have dimensions table in some jdbc database
> (product,location,time,customer) and a fact/summarized data in hdfs file
> (salesfact).
>
You would create one DBStorage and call it with some name say - mydb. The
tables product, location, time, customer will be associated with DBStorage.
And Create one HDFSStorage and call it with a name, say - mycluster.
SalesFact will be associated with HDFSStorage. See example definitions here
-
https://github.com/apache/incubator-lens/tree/master/lens-examples/src/main/resources

But if you have dimtables in DBStorage and facts in HDFSStorage, there is
no driver which can talk to both those storages. You might have to
replicate dimtables into HDFS, sothat Hive can read them.


> I want to create a cube in Apache lens which can slice the sales per
> product,customer and location. How can I create dimension,facts reference
> tables in lens for the above mentioned scenario and if we need to create
> storage for each separately what is the purpose.
>
> Hope the above helped. If not, ask more questions


> Regards,
> Srinivasan Hariharan
>
>
>
>
>

Mime
View raw message