"msmbarabino@virgilio.it" <msmbarabino@virgilio.it> writes:
> The test table has 36 records and each bytea field
> contains a 250KB object. A (SELECT byteafield FROM table) takes more
> than 1 minute to execute.
Hm, it takes about 4 seconds here, on a fairly old and slow machine:
$ time psql -A -c "select * from file" regression | wc
38 33699 26653873
real 0m4.00s
user 0m1.37s
sys 0m0.62s
In psql the trick is to use unaligned display mode, else psql itself
eats a whole lot of time trying to nicely format those 250KB rows:
$ time psql -c "select * from file" regression | wc
40 33700 28134681
real 0m57.10s
user 0m53.55s
sys 0m0.67s
I surmise that your problem is likewise on the client side, but you
weren't very specific about what client code you were using. Look
for bottlenecks associated with processing of very wide rows...
regards, tom lane
