In my qliksence application i have around 80 tables.basically this is a conversion project what ever they implemented in their tool we need to implement the same here but the problem is the tables having huge like 17cr records in almost 30 tables among that 15 table are not having anydatefileld and filters.if i load the data app size is exceeding 250mb .so please help me out how to create the data model.
In my fact table i have date field as a filter
and also few tables are having more than 5 date fileds
With the information provided, I can only make overall recommendations. 80 tables is a lot. The most complex model I have built has about 20 tables (including island tables for dimensions and control purposes) and this was integrating data from about 25 different sources. Having said that, 3 million rows and 250 MB is not a very large data set.
You need to work towards a star structure with a single central fact table and dimension tables linked to the fact table. You could also collapse small dimensions (eg containing only some ID/Name data) into the fact to simplify the structure.