[BUGS] Bad pg_stats.correlation with repeated keys.

2013-08-15 Thread Ralph Loader
attempt to actually match the on disk data order. [For my use case, repeated keys insert after existing would be right, I am not sure if that is always true.] Cheers, Ralph Loader. PS. A tangentally related comment : bitmap scans achieve good ordering for the data reads, but not necessarily for

[BUGS] Bad pg_stats.correlation with repeated keys.

2013-08-16 Thread Ralph Loader
attempt to actually match the on disk order. [For my use case, repeated keys insert after existing would be right, I am not sure if that is always true.] Cheers, Ralph Loader. PS. A tangentally related comment : bitmap scans achieve good ordering for the data reads, but not necessarily for the i