I'm simply referring to literature (like the intro Ramakrishnan & Gehrke).
I just know that Oracle an Mysql actually do have them too and use it
without those current implementation specific restrictions in
Postgres.
IMHO by design Hash Index (e.g. linear hashing) work best when:
1. only equal (=) tests are used (on whole values)
2. columns (key values) have very-high cardinality
And ideally but not necessarily when index values do not change and
number of rows are known ahead of time (avoiding O(N) worst case - but
there are approaches to chaining with dynamic resizing).
I just collected this to encourage ourselves that enhancing hash
indexes could be worthwhile.
Stefan
2011/9/18 Kevin Grittner <Kevin.Grittner@wicourts.gov>:
> Stefan Keller wrote:
>
>> It's hard for me to imagine that btree is superior for all the
>> issues mentioned before.
>
> It would be great if you could show a benchmark technique which shows
> otherwise.
>
> -Kevin
>