seperate qvds for alldimentions and facts

[sonysree88]

Hi All,

Can you any one please help for this question.

Its need urgent to me implement in the work area

i need to add all the dimension tables in one qvd and all the fact tables in one qvd based on the common column which will be the ROWID or another one we need to join both the qvds itself ,its the data model which we need to create and we should use this one in presentation layer to create the dashboard or report.

[stabben23]

Hi,

sounds to me like a stupid idea, why not keep the Dimension in their own qvd's?

Build Your datamodel and use binary load or create the Whole qvw to and qvd as Rob expain here.

Storing a Data Model in a Single QVD | Qlikview Cookbook

[MK9885]

You cannot create one QVD for separate dimensions (unless you join/concatenate...) them. Depending on fields/data it can be done but if you quite a few dimension tables then it's not a good idea to join them all.

If you just need to get Data model in your Source file then follow the below procedures.

Ex:

If you have 10 dimensions Tables

Create QVD's for 10 Dimensions Tables

And for Fact Table create a QVD as well.

Use all those 10+1 (Dimension + Fact) Table QVD in one QVW file (Name it s Data Model)

Create a new QVW and do binary load of your Data Model QVW and it will be your main/source file.

Hence you can create visualizations in front end.

You just need to do binary load to get all Dim & Fact Table.

[awhitfield]

Why?

[sonysree88]

when we are using the binary load its taking lot of time thats what we got the guidelines from archtects like to create seperate qvds

1 is for to integreate all the dimensions in 1 qvd

2 is for all the measures in 1 qvd

3 is parter-ref qvd

we can combine all the 2 qvds based on fact id which we can create based on row-id

[jonathandienst]

I would not support the idea of concatenating or joining all the dimensions into a single table (which is what you would need to do). Unless your data model is very small (no more than a few thousand rows in the fact table and a few hundred rows and a small when you combine all the dimensions), this is likely to perform badly and likely to require more complex expressions.

Loading 10 dimensions from 10 QVDs is a quick and simple operation, so I cannot see a good reason for your requirements.

[jonathandienst]

And if the binary load is taking a long time, you are unlikely to save significant reload time by following your proposal as well as possibly severely affecting front end performance (which is usually more important than reload time).

[sonysree88]

Fact QVD                 : Contains only key figures/Measures, and unique ID

KEY QVD                : Contains only dimensions, and unique ID.

Partner Ref QVD           : Contains SAP Customer ID, Partner Role and unique ID

---->Unique ID  identifies unique row in fact table, and will be link among these three files.