Hello, Nathan. I update from 2.0.5 to 3.1.1. On our project we use Postgres and we must use many preConditions, and perfomance of preconditions is very important.
I will run some benchmark
I took next result (liquibase 2.0 vs 3.1):
2 min 21 sec fk_by_name
2 min 19 sec index_by_name
2 min 18 sec index_by_name_and_table
2 min 17 sec pk_by_name
2 min 19 sec pk_by_name_and_table
2 min 19 sec seq_by_name
53 sec fk_by_name
2 min 33 sec index_by_name
1 min 31 sec index_by_name_and_table
34 sec pk_by_name
33 sec pk_by_name_and_table
9 sec seq_by_name
It is very bad results too. We wrote some stored functions and use them like:
But this way not so beautiful. Can we use database specific methods for preconditions by dbms parameter?
changeSet was run on schema with about 1000 tables
Thanks, I'll take a look
Snapshot performance is improving, but needs more work so leaving open until 3.3.
This should be fixed along with changes in