Post reply

A big table or subset of tables (Datawarehouse q.)

  • September 7, 2006 at 6:40 am

    #115337

    I am working on a dataware house problem. a excel ark each time required will be dumped into datawarehouse.

    Number of fields are 80. is it suitable to copy everything in one table in sql server

    or create different tables based on partial concepts and split data into these tables ?

    whic is best for database performance. I mean that many tables with small number of columns are better then just one table with all columns.

    is there any difference when we make a view on one table solution or make a view by joining all tables.

    how will be sql query performance between these two models ?

     

     

    with my regards

     

  • September 11, 2006 at 4:01 am

    #659582

    The easiest answer would be:

    Read up on RDBMS and how they work and design your tables from there. If performance is a problem denormalize a little...

    Then if you want to put a view on your tables read up on partitioning  (check-constraints if you are using SQL Server 2000). Also interesting for performance, if you are using views, is SQL Server join elimination (kind of implicit vertical partitioning).

    Happy reading

Viewing 2 posts - 1 through 1 (of 1 total)

You must be logged in to reply to this topic. Login to reply