Support UUID columns across more databases #46
Conversation
Postgres returns a concrete `java.util.UUID` where MySQL returns a byte array. This should improve support for any databases that support a native UUID column type.
| case x: Array[Byte] => x | ||
| case x: Blob => x.getBytes(0, x.length.toInt) | ||
| case x: Clob => x.getSubString(1, x.length.asInstanceOf[Int]).getBytes | ||
| case x: String => x.toCharArray.map(_.toByte) |
There was a problem hiding this comment.
maybe if it is a string it should be treated a s dash-seperated hex-encoded UUID?
There was a problem hiding this comment.
You're probably right, but that would be a change that breaks the API. This code is taken from the implementation of byteArrayOption which is what uuid used to use.
|
Thoughts: (1) This is why I don't like (2) I'd rather we stick to the JDBC types. We got in trouble because we had a bright idea (UUIDs) that didn't always work. If we have logic we'd like to add to the standard JDBC interface, it should be in driver-specific projects. This PR is good; it can only help in our current situation. In relate 3, I'd like fix these mistakes. |
Postgres returns a concrete
java.util.UUIDwhere MySQL returns abyte array. This should improve support for any databases that support
a native UUID column type.
fixes #45