[sqlalchemy] Re: Advice for Implementing a ShortUUID Column Type?

2017-05-20 Thread Jonathan Vanasco
the postgres trick is to use partial index on the column to improve the query speed. 1. create a substring index on the table, lets say 5 characters wide: CREATE INDEX speed_shortuuid ON table(substr(shortuuid, 0, 5) 2. have all queries include a match against that index, calculating the

[sqlalchemy] Re: Advice for Implementing a ShortUUID Column Type?

2017-05-19 Thread Colton Allen
I do not know the optimization trick. I'd be interested to know! It would be nice to not have to translate to and from the UUID type. On Friday, May 19, 2017 at 4:04:55 PM UTC-7, Jonathan Vanasco wrote: > > side question - have you done any tests on how the UUID type queries as > your

[sqlalchemy] Re: Advice for Implementing a ShortUUID Column Type?

2017-05-19 Thread Jonathan Vanasco
side question - have you done any tests on how the UUID type queries as your database scales? It may make sense to do the shortuuid as a text field, which you can put a substring index on. if you don't know that postgresql optimization trick, I can explain it further. -- SQLAlchemy - The