Hannu Krosing <hannu@tm.ee> writes:
> What are the performance implications - is it 1.01, 10 or 1000 times
> slower than accessing the same file from fs for files in the range of
> typical Office documents (0.2-20M) ?
I think that's a meaningless question unless you specify a particular
usage pattern --- but in any case, I doubt anyone has tried to
characterize it that way. Feel free to run some experiments and tell
us about your results...
> Will updating one non-toasted field in a tuple copy the toasted one
> as well, or is only the reference copied ?
It copies the reference, see toast_insert_or_update in
src/backend/access/heap/tuptoaster.c.
Speaking of which, though, it looks like an update or insert will
forcibly uncompress (and later recompress) a compressed-in-line datum,
which seems like a waste of cycles to me. Jan, shouldn't the test for
VARATT_IS_EXTENDED at line 357 instead read VARATT_IS_EXTERNAL?
regards, tom lane