Database Developement Theory. One table with 500 columns vs 1000 individual Tables.
Posted on 2016-10-17
Thank you all in advance!
We are importing legacy data from 100's of tables. Those tables have very similar columns. In fact 35% or so of the column are the same columns that are actually capturing the same data type. Quick origin story on that, the legacy application has one table for each webform. Thus, over time, the deployment of 100s of webforms has lead to 100's of tables.
I want to clean this up. If I can. To that end, I read Sql Server 2016 can have a max of 2,147,483,647 objects (UDFs, stored procs, tables views, etc). Thus, I don't think we will run into issues on the sql side continuing to make a table for each webform. However, I get the feeling the JOIN to bring all the webform data together would be...epic.
I could spend some time consolidating all those tables into one, standardizing column headings to eliminate redundant columns and make unique ones where necessary. However, that would lead to a table with about 500 columns.
My feeling is that proper Normalization of the tables would produce one table in lieu of the current multi table structure. Was just looking for your thoughts before boiling down those tables.
thanks again in advance!