I have in my project liquibase set up and all is working fine. There is initial xml files, made one year ago and a lot of change sets that are made meanwhile.
I run all change-sets for test executions. That mean I run initial change set and also all differential scripts(new columns, changed types). I want to reduce time of this execution. I was thinking about this:
1. on existing database run generateChangeLog for both structure and data (in this way I will get much less code, because there way a lot of changesmeanwhile, eg. add column table.column1, change type of table.column1, delete column,...... This mean that I will have much less change sets)
2. Instead of old files I want to use generated ones.
I have stable environments where old change sets have been executed. This mean than table databasechangelog is already populated with old values. (values from old files)
Now, I want to execute new change set what will be written starting now in old environments. This will not work because for example: I have table A in my DB, but change set I am executing has: create table A;
What is best way to solve this problem?
In short: I have databasechangelog populated, but I have to re-create all files, and I want to be able to run re-created change sets for test, but also to be able to execute same xml files for future sets.