Is it mandatory to maintain a star schema/ snow flake schema at QlikView end I.E. Have all the measures defined in the fact surrounded by dimensions if you have a lot of data sources or can you just create the data Model based on common fields. Do you need to follow a mandatory rule while the model is getting created based on these common fields
It is not necessary that you create Star schema every time but try to create start schema because it is easy to use and when you keep maintaining and adding new tables into your data model then it is easy to maintain rather than snowflake schema because in snowflake you get the challenge of the slowly changing dimensions.
I don't think there is any mandatory rule to create a data model.
Star schema is considered as best way but sometimes snow flake can also be done (depending on how the data is?)
Having most of the possible measures defined in script is good way to increase the performance ( Ex- basic If conditions to create a new dimensions).
Removing NULL values if required from the back end.
Using Where conditions at the end of the table to bring only filtered and required data.
Ex- Where Year > '2015'; <<< this would only bring data from 2015 and not old data.
Using SQL query to filter the data from data base but not bringing millions of rows which are not required for our raw extract. same where conditions can be used or joins/mapping etc.
