Multiple connections in one universe

building semantic layer
1

HI All,

I have scenrio like below

I need to access data in one universe with two differnt connecions and accessing user will remain same users for connection1
and 2.

example:

Connection - XYZ is pulling data from A b c table targetted users are user1 and user2 and user3

Connection - KLM is pulling data from A1 b1 and c1 table targetted users are user1 and user2 and user3

you might ask one question here since both tables are in same schema why dont you create one universe…
my requirement is like this i need to access both are with separate connection :frowning:

I have two option for this

  1. create 2 universe and make and link each other and make one as master universe - which i am not intrested
    2.create 2 separate groups in cmc and assign the connection to the each group i am looking for this ?
    but i need some more information to achive this because i haven’t impmented this

and alos if there is any best option please provide me

env:BOxi3.1
db: sql server

pavan_au143 (BOB member since 2007-07-25)

2

I’m not sure that I understand your requirements…

If your tables are in a single schema, and you want one universe, use one connection. I don’t understand why you would use 2 connections, especially as the users are the same.

  1. Linked universes: AFAIK, the two linked universes use the same connection.

  2. I don’t know what you want to achieve by using the CMC: this is useful to giver a different connection to different users.

KFonMurphi :fr: (BOB member since 2007-10-16)

3

Hi thanks for your replay,

I have two diffent set of data which are maintaing in two differnt tables withing the sechema

like example 1.current data( transacational load ) and 2. history data ( from yestr day to 2 year) now users want data in one area duw to some performnce issue we are not hitting history tables we r tranforming history data to another server and creating universe from there.

and finally link these two…so i wanted to avoid separate universe if we have any option available which solve my scenrio here?

please let me know if you need more information

thanks

pavan_au143 (BOB member since 2007-07-25)

4

One single connection any way. I cannot see how having two connections would help.

I would try a big derived table with UNION ALL of your fact tables. I don’t know how it will behave if you query only the current data, perhaps awfully. Perhaps a view at databse level would be quicker.

Another solution is to create two contexts. Supposing that you have a clean star schema with common dimensions, and just a historical fact table, and a current fact table: create measures on both tables, eg. [Amount (history)] an [Amount (current)]. If the user wants history AND current amounts, he uses them in the same query, Webi makes 2 SQL queries, and you add both measures in the report.

The cleanest way is at dabatabase level, with a single table, perhaps with partitioning, but it depends on your level of control, the latency, ETL tools…

KFonMurphi :fr: (BOB member since 2007-10-16)

5

Go with option 2, give it a try in your dev environment. It works fine.

6

If the group of users is the same for each group, how will you override the connection?

Something along these lines makes sense to me, although the OPs original explanation of his situation is a little unclear.

Mak 1 :uk: (BOB member since 2005-01-06)