On Sat, Jan 12, 2008 at 05:47:30PM +0000, Simon Riggs wrote: > On Sat, 2008-01-12 at 01:59 +0100, Gavin Sherry wrote: > > The syntax is half the problem, performance is the other. > > The syntax looks great to me, but I think it is about 5% of the problem, > maybe less. I don't really have any questions about the syntax, but I > may have thoughts when the implementation details emerge.
Yes, that's for another thread. Since the discussion was abot using grammar to control partitions I wanted to get some grammar out. More details on other stuff soon. > > I'm not sure you'll be able to use PARTITION BY since its part of the > SQL Standard for Windowed grouping, which we do hope to implement one > day. It will be confusing to have two completely separate meanings for > the one phrase in our grammar. I think it's fine. It doesn't cause conflicts in the grammar (in fact, the Greenplum grammar implements both meanings right now with no confusion). Thanks, Gavin ---------------------------(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