Is it possible to have a persistent psql connection in bash? - database

Question

I have a very large bash script and I make thousands of psql queries throughout it which I suspect is slowing it down. I have created this MWE which proves my theory. I am assuming this slow down is because of having to connect to the db repeatedly. Is there a way to remain connected to psql in bash?
#!/bin/bash
DB_NAME=testdb
DB_USER=user1 #UPDATE THIS
#DROP DB
SQL_QUERY="DROP DATABASE IF EXISTS $DB_NAME;"
echo $SQL_QUERY | sudo -u $DB_USER psql >/dev/null
#CREATE DB
SQL_QUERY="CREATE DATABASE $DB_NAME;"
echo $SQL_QUERY | sudo -u $DB_USER psql >/dev/null
#CREATE TABLE
SQL_QUERY="CREATE TABLE foo
(
id BIGSERIAL PRIMARY KEY,
problems INTEGER
);"
echo $SQL_QUERY | sudo -u $DB_USER psql -d $DB_NAME >/dev/null
TARGET_ITERATIONS=1000
#MULTIPLE DB CALLS
SQL_QUERY="INSERT INTO foo
(
problems
)
VALUES
(
99
);"
START_TIME=$(date +%s)
for ITERATION in $(seq $TARGET_ITERATIONS)
do
echo $SQL_QUERY | sudo -u $DB_USER psql -d $DB_NAME >/dev/null
done
STOP_TIME=$(date +%s)
echo "Multiple Call Duration (s): $((STOP_TIME-START_TIME))"
#Single db call
SQL_QUERY=""
for ITERATION in $(seq $TARGET_ITERATIONS)
do
SQL_QUERY+="INSERT INTO foo
(
problems
)
VALUES
(
99
);"
done
START_TIME=$(date +%s)
echo $SQL_QUERY | sudo -u postgres psql -d $DB_NAME >/dev/null
STOP_TIME=$(date +%s)
echo "Single Call Duration (s): $((STOP_TIME-START_TIME))"
Output:
$ ./test_psql.sh
Multiple Call Duration (s): 64
Single Call Duration (s): 12

Answer 1

Per your comment, try running this:
#/bin/bash
mkfifo /tmp/mypipe
chmod a+r /tmp/mypipe
sleep 10000 > /tmp/mypipe &
psql -f /tmp/mypipe &
echo "\pset pager off" > /tmp/mypipe
echo "select count(*) from information_schema.columns;" > /tmp/mypipe
echo "select count(*) from information_schema.columns;" > /tmp/mypipe
sleep 3
echo "select count(*) from information_schema.columns;" > /tmp/mypipe
echo "\q" > /tmp/mypipe
rm /tmp/mypipe