Joining Tables in a Data Model

[Previous Member]

Hi, 

 

I am a little bit confused, I started to work on Board but the last version is quite different from the older one. 

 

To load a cube (number of employees per year), I need to join tables in my data model. 
I can't find how to joint, I suppose it's in the Data Reader but I do not undersand how to do so. 

 

I find this on Board Manual : 

“If any JOIN is needed, it can be defined under the JOIN tab, clicking on the ADD JOIN button and then selecting the fields to join from the 2 drop down lists. A where condition can be defined manually under the WHERE tab.”

=> There is no tab or button. 

 

Can you help me ? Or maybe give me some examples, it would be very nice, 

 

 

Thank you by advance for ur help, 

 

Have a nice day, 

Paul Wyatt

Hi C Tnu,

 

You're probably in a hurryr so I'll show you how to construct a join to build a 1D cube and hope that I cover the step you were looking for.  Get back to me if I have not helped.

 

To join 2 tables to build a 1D cube

 

Step 1.  Construct your cube structure within BOARD

 

Step 2.  Go to the [DATA READER] AND select [+ DATA READER] to create a new one.

 

Step 3.  Construct your cube as needed, using any fields from any tables required

Step 4.  Select [JOIN]

Step 5. Select [ADD JOIN].  Use the 'down' arrows to list the fields of both tables to join from.

Step 6.  Your JOIN has been created as an INNER Join and will run.Read the next step to change JOIN type

Step 7.  Switch to [MANUAL] to overtype the SQL Script for other JOIN types.

 

Hope this helps

 

Regards,

 

Paul

[Previous Member]

Thank you so much Paul Wyatt - Avison Young UK for your quick reply ! 

 

Actually, I can't find this 'JOIN' area.

I maybe forgot specifying that I'm on Text file source, I use CSV files : 

 

Maybe we can work only with one file but it seems strange, isn't it ? 

 

 

Thanks by advance for your help, 

 

 

Best regards, 

Paul Wyatt

Hi C Tnu,

 

I'm not aware that you can join on txt files within the Reader.  You may be able to achieve this with DATA Fasttrack, but I could not see anything to achieve this in there either. 

 

My advise would be to create an intermediate structure to bring both sets of data into their own cubes, followed by using a dataflow to merge them as required.

 

I can only advise you of what I have done in these circumstances.  I don't know how complex your data is and so cannot say for sure if this approach would be appropriate, Sorry

[Previous Member]

Ok, thnak you very much !

 

I'll try DataFastTrack. This is for a test so I took several complex files to train ! 

 

Bye, 

Paul Wyatt

OK C Tnu, good luck in your test.  I'd be very interested to learn of any alternative and creative ways of joining datasets so please, let me know what you find?

 

Kind regards,

 

Paul

[Previous Member]

Of course, I'll keep you informed as soon as possible. 

 

Kind regards,