Problems with 'diff' and primary key names

General Discussion
1

Hi

I’m new to liquibase and have the following issue: We’re comparing two db schemas with the diff functionality of liquibase to make sure no db change was forgotten during development. This works great - only with primary keys, I get a lot of ‘phantom differences’ due to different primary key names. Primary keys have been defined anonymously in the ‘CREATE TABLE’ statement. H2 obviously names them with some internal name that trigger a diff.

Is there a mode where liquibase diff ignores the names and only checks the attribute set of a primary key? Or do I have to name all my primary keys? Is there a way to name a primary key within the create table statement?

Thanks!
Remo

Example:
…
Missing Indexes:
     …
     PRIMARY_KEY_38 on INSTALLATION(INSTALLATION_ID)

Unexpected Indexes:
     …
     PRIMARY_KEY_8782E on INSTALLATION(INSTALLATION_ID)


2

Not a good way in 2.x. The diff/snapshot/comparison functionality is the main focus of 3.0 which should help with this sort of thing. 


For now, you’ll just have understand from the diff report that the name differences are not meaningful to you.


Nathan