Dear group, Thank you for your feedback in this matter. The feedback and conclusions I got are these:
1. Not many users would immediately care about such a feature. It is certainly not an urgent one 2. Users closer to SQL than to Java might appreciate this feature, if properly highlighted 3. It is quite complex to fit this feature into the existing record type-hierarchy Given that the current record type-hierarchy might need some re-thinking, it is probably unwise to introduce such a change in a minor release (i.e. 3.1). Nonetheless, generic updatability could be a very nice addition to jOOQ, so I will definitely schedule this as an enhancement for 4.0 Cheers Lukas 2013/6/5 Lukas Eder <[email protected]> > Hello, > > Thanks for the feedback and sorry for the delay > > 2013/5/27 Durchholz, Joachim <[email protected]> > > > #1846 >> >> I like the features. >> >> I dislike the multitude of modes that the resultset could be in. You >> can't infer its contract from the type anymore, you have to inspect its >> configuration as well. >> >> The knee-jerk approach to solve that would be a separate subclass per >> mode (assuming you can't change the mode after creating the resultset). >> I don't know whether that would create problems elsewhere. >> > > I currently want to move in the opposite direction, actually. I think that > having reserved a type for updatability (UpdatableRecord vs. Record) was a > mistake in early jOOQ days. There are several reasons why I have come to > this conclusion: > > - It is hard to assess updatability of a record in SQL (we've discussed > this before). It isn't strictly related to the presence of a primary / > unique key value in a record. Most databases also provide ROWID, OID, or > other pseudo-columns, although they usually hide them from "upstream" APIs, > such as JDBC. > - Updatable via PK is not strictly the same as Updatable via ROWID / JDBC > updateXXX() methods. Merging this new feature in the existing > UpdatableRecord is a pain, creating a new JDBCUpdatableRecord is an even > bigger pain :-) > - The "early jOOQ" had too many types in its type hierarchy in general. > Not everything needs to be specialised all the time. > > >> >> > #2265 >> >> Can't say much about that one, I'm unsure what contracts are affected. >> However, this: >> >> > making all records updatable if fetched accordingly. >> >> sounds like that may be losing type information, i.e. that Record >> couldn't update to the database before and now it can; that's a loss of >> information available to the code maintainer, because now he has to infer >> the difference from data flow and can't rely on types anymore. >> I may be misinterpreting this though. >> > > You're interpreting it correctly. > > Cheers > Lukas > -- You received this message because you are subscribed to the Google Groups "jOOQ User Group" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
