Reference documentation for column type conversion and support

I apologize if this has been answered already, but I’ve been searching for a reference of changelog column type to database-specific type and can’t find any.

I really just want to know the types that are mostly database independent.  My project requires support for Oracle and Postgres and I want to make sure our changelogs being developed are compatible with both.

This is all I have found on database-independent types:

<span data types will be converted to the correct database implementation: * BOOLEAN * CURRENCY * UUID * CLOB * BLOB * DATE * DATETIME * TIME * BIGINT Also, specifying a java.sql.Types.* type will be converted to the correct type as well. If needed, precision can be included. * java.sql.Types.TIMESTAMP * java.sql.Types.VARCHAR(255)

I believe there used to be a table in the addcolumn documentation that gave me the information I’m looking for on the old web site.  Something like this:

Type      | Oracle   | Postgres | etc…


Number  | Number | Not Sup  | …

If there is a reference like this, can someone point me to it?  Thanks

I’ve been looking for the same thing. 

Closest thing I’ve found is a pointer to the source code:

And Javadoc for the SQL types:

My guess is that the conversion map is complex and in constant flux, so a static reference may not be in sync with the version of the code someone is using. Still, a table like you describe would be helpful for each release version.  

Perhaps there is a wiki we could create such a page on? 

Or maybe the type-converter classes themselves could be queried or modified to build such a table? 

My plan was to use the type converter classes to auto-gen the documentation, but I haven’t quite made it there yet. It has tended to be in a bit of flux which is part of the reason it hasn’t been well documented. I’m hoping to put some final touches in 3.1 and really stabilize and document it.


Thanks for the reply and heads up, Nathan. I’ve been digging into more and more of the cmd line and other options in the past couple weeks, and the more I wrap my head around it, the more I’m impressed with Liquibase. 

Having a matrix of conversions would be a big help in grokking what SQL will be created from generic definitions. 

Anything I could do to help? 

Oops. Wasn’t logged in, posted that as a Guest. Anything anyone could do to help?