Dear all, I've written a bash script that looks like the one below: #!/bin/bash DBuser='root' DBname='test' psql -q --username=$DBuser --dbname=$DBname -c "prepare test_statement (integer) as insert into tbl_test (col_test) values (\\$1)"; tail -f /root/testfile | while read a; do psql -q --username=$DBuser --dbname=$DBname -c "execute test_statement ($a)"; done; psql -q --username=$DBuser --dbname=$DBname -c "deallocate test_statement"; Note that this is very stripped version of the real script, but it gives the same errors: ERROR: prepared statement "test_statement" does not exist I persume that this is caused because of the individual statements, each using their own session. But is there a way to avoid this? In reality the statement that is prepared is much more complex, and the files that are processed are rather big. I hoped to gain more performance by preparing the statement. Your sincerely, Aarjan Langereis Ps. The bash script is ran on an Fedora Core 3 machine using PostgreSQL 8.1.0