I’m always more than happy to see how dbdesc users customize dbdesc templates for their own needs. Last week, Jasper Verweij posted a comment sharing his idea of building tag clouds using database objects.

It’s indeed a very useful idea because it shows which objects are “more important” on your database. After a couple of email exchanges he was able to put in place a tagcloud of referenced tables. That is, the more referenced a table is, the bigger is its name. This picture shows the effect:

This is just a proof of concept. He is working on more tagclouds… In his own words:

The next thing I’ll try to make, is a flexible function to make the tagcloud functionality available for various items of a database, like:
- Tagcloud by UDF dependencies
- Tagcloud by SP dependencies
- Tagcloud Tables by View
- Tagcloud Tables by Table Size
- Tagcloud Tables by Row count
- Tagcloud Tables by Index Size
-Tagcloud by data type (which datatypes are used the most)
etc.

After this setup another trick could be tried: combination of tagclouds. A big table (with loads of rows) with a primary key which used a lot as foreign key might be more important than a table which has 5 rows and 1 foreign key.

If anyone is interested in the code to add this tag clouds to dbdesc templates just drop me an email.