Peter Geoghegan <pg@bowt.ie> writes:
> On Mon, Sep 18, 2017 at 7:31 AM, Neto pr <netopr9@gmail.com> wrote:
>> In my example, the values of fast_root, fast_root are equal to root, level,
>> I believe that due to the newly created index and no delete operations
>> occurred in the table.
> Fast root and true root will probably never be different, even when
> there are many deletions, including page deletions by VACUUM. As I
> understand it, the fast root thing is for a fairly rare, though still
> important edge case. It's a way of working around the fact that a
> B-Tree can never become shorter due to the locking protocols not
> allowing it. We can instead just pretend that it's shorter, knowing
> that upper levels don't contain useful information.
My (vague) recollection is that it's actually useful in cases where the
live key-space constantly migrates to the right, so that the original
upper-level key splits would become impossibly unbalanced. This isn't
all that unusual a situation; consider timestamp keys for instance,
in a table where old data gets flushed regularly.
regards, tom lane
--
Sent via pgsql-performance mailing list (pgsql-performance@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-performance
