Kris Jurka on 12/02/08 16:40, wrote:
On Tue, 12 Feb 2008, Adam Hardy wrote:
Because this rules out certain important features of the JPA
framework such as native SQL queries, you may want to prioritize this
issue. I will have to use Oracle or mySQL until PostgreSQL can
rectify things.
I wouldn't expect postgresql to change anytime soon. Could you
explain in more detail what you can't use and why the JPA side
couldn't be fixed to support postgresql?
You could also consider always quoting all identifiers in both
database creation scripts and queries. This guarantees that the
database won't mess with the case of the object although it can be a
pain to type if you writing a lot of queries by hand.
Well, JPA is just a standard, so a 'JPA fix' would require each JPA
provider (Toplink, JDO, OpenJPA, Hibernate etc) to rectify their
postgresql-specific dialect / SQL syntax builder.
This issue arises using PostgreSQL with a schema name when running
native SQL queries.
That means I write pure SQL in my JPA mapping files. It's a common
practice.
Using a schema name is also a common practice.
However I will, at some stage in the future, code a test batch, which
tests basic CRUD functionality and OR mapping across different
combinations of JPA providers and DB vendors. So I will keep the bug
reference handy and let you know how many JPA providers suffer this
issue once I run the test.
Best regards
Adam
---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match