Suggestions for schema modification.

[EvanCarroll]

There are a lot of suggestions I have regarding modifications to the Chado schema. How do I best make those suggestions? Is there an automated build system?

You have a lot of composite natural keys where the first value on the key is also indexed. There is no reason to ever do that.

• featureloc_idx1
• analysisfeature_idx1
• synonym_idx1
• dbxref_idx1

Amongst others are good examples of where you can save GBs of ram and disk space by eliminating useless indexes.

This kind of stuff,

• the default of ('now'::text)::timestamp(6) with time zone

Is just a slower method of saying DEFAULT now().

boxquery and boxquery are a costumed rolled and much slower method of doing operations on an intrange with a gist index. Problem being PostgreSQL provides a much more featurefull standardized and optimized method with the int4range type.

Storing the md5 as character(32) is also a bad idea, those should be stored as UUID which would save a ton of space in index and disk.