PostgreSQL
Login to PSQL
export PGPASSWORD='p@55w0rd'
# [email protected]:~ #
psql -U postgres
#psql (10.15 (Ubuntu 10.15-0ubuntu0.18.04.1))
#Type "help" for help.
# [email protected]:~ #
su - postgres
# Last login: Tue Jan 01 12:00:00 MYT 2013 on pts/0
# -bash-4.2$
psql
# psql (9.6.8)
# Type "help" for help.
#
# postgres=#
Login Error
# [email protected]:~ #
su - postgres psql
# Last login: Tue Jan 01 12:00:00 MYT 2013 on pts/0
# /bin/createuser: /bin/createuser: cannot execute binary file
Creating User
# [email protected]:~ #
su - postgres
# Last login: Tue Jan 01 12:00:00 MYT 2013 on pts/0
# -bash-4.2$
createuser root
# createuser: creation of new role failed: ERROR: role "root" already exists
# ALTER USER root WITH SUPERUSER;
Creating DB
Until creating root database it's always show error could not change directory to "/root": Permission denied. To overcome this error we should need to create a root database. only for avoid error while login using root user.
# [email protected]:~ #
su - postgres
# Last login: Tue Jan 01 12:00:00 MYT 2013 on pts/0
# -bash-4.2$
createdb root
Similarly you can create a database as same as your frequently used OS User. In this case no need password to login. It will authenticated from OS. That's means you can login to your PostgreSQL user using OS Authentication.
Grant Privileges
# postgres=#
grant all privileges on database root to root;
# GRANT
# postgres=#
grant all privileges on database postgres to root;
# GRANT
Logout PSQL
# postgres=#
\q
# -bash-4.2$
Meta Data
-- show databases
SELECT datname FROM pg_database;
-- show tables
SELECT * FROM pg_catalog.pg_tables
WHERE schemaname NOT IN ('pg_catalog','information_schema');
-- describe tables
SELECT table_name, column_name, data_type
FROM information_schema.columns WHERE table_name = 'my_table';
Array Query
SELECT dependencies||'"Running Fiscal Year"'
FROM scope WHERE name = 'Fiscal Year Config';
JSON Query
{
"type": "Reseller",
"commission": {
"type": "Percentage",
"value": "25"
}
}
UPDATE seller
SET seller_properties = jsonb_set(seller_properties, '{type}', '"Retailer"')
WHERE id = 1000;
--
UPDATE seller
SET seller_properties = jsonb_set(seller_properties, '{commission, type}', '"Bonus"')
WHERE id = 1000 AND seller_properties ->> 'type' = 'Reseller';
|
JSON SELECT id, name FROM json_to_record (
'{"id":1,"name":"Academia"}'::json
) AS concern(id int, name text);
SELECT * FROM json_array_elements_text('[
{"id":1,"name":"Academia"},
{"id":2,"name":"Agronomy"},
{"id":3,"name":"Software"}
]'::json);
SELECT id, name FROM json_array_elements_text('[
{"id": 1, "name": "Academia"},
{"id": 2, "name": "Agronomy"},
{"id": 3, "name": "Software"}
]'::json) AS concerns_text,
json_to_record(concerns_text::json)
AS concerns_json(id int, name text);
SELECT id, name FROM json_array_elements('[
{"id": 1, "name": "Academia"},
{"id": 2, "name": "Agronomy"},
{"id": 3, "name": "Software"}
]'::json) AS concerns_array,
json_to_record(concerns_array)
AS concerns_rows(id int, name text);
SELECT id, name FROM json_to_recordset('{
"data": [{
"concerns": {"id": 1, "name": "Academia"}
}, {
"concerns": {"id": 2, "name": "Agronomy"}
}, {
"concerns": {"id": 3, "name": "Software"}
}]
}'::json -> 'data') as r(concerns json),
json_to_record(concerns) AS concern(id int, name text);
|
JSONB SELECT id, name FROM jsonb_to_record (
'{"id":1,"name":"Academia"}'::jsonb
) AS concern(id int, name text);
SELECT * FROM jsonb_array_elements_text('[
{"id":1,"name":"Academia"},
{"id":2,"name":"Agronomy"},
{"id":3,"name":"Software"}
]'::jsonb);
SELECT id, name FROM jsonb_array_elements_text('[
{"id": 1, "name": "Academia"},
{"id": 2, "name": "Agronomy"},
{"id": 3, "name": "Software"}
]'::jsonb) AS concerns_text,
jsonb_to_record(concerns_text::jsonb)
AS concerns_json(id int, name text);
SELECT id, name FROM jsonb_array_elements('[
{"id": 1, "name": "Academia"},
{"id": 2, "name": "Agronomy"},
{"id": 3, "name": "Software"}
]'::jsonb) AS concerns_array,
jsonb_to_record(concerns_array)
AS concerns_rows(id int, name text);
SELECT id, name FROM jsonb_to_recordset('{
"data": [{
"concerns": {"id": 1, "name": "Academia"}
}, {
"concerns": {"id": 2, "name": "Agronomy"}
}, {
"concerns": {"id": 3, "name": "Software"}
}]
}'::jsonb -> 'data') as r(concerns jsonb),
jsonb_to_record(concerns) AS concern(id int, name text);
|
Result id | name ----+---------- 1 | Academia value
-------------------------------
{"id": 1, "name": "Academia"}
{"id": 2, "name": "Agronomy"}
{"id": 3, "name": "Software"}
id | name ----+---------- 1 | Academia 2 | Agronomy 3 | Software (3 rows) id | name ----+---------- 1 | Academia 2 | Agronomy 3 | Software (3 rows) id | name ----+---------- 1 | Academia 2 | Agronomy 3 | Software (3 rows) |
Delete Join
-- USING is not an ANSI standard not supported by others RDMBS
-- it's better to use sub query rather than USING for join
DELETE FROM product p USING category c
WHERE p.category_id = c.id AND
c.name = 'Grocery';
Sequence
CREATE SEQUENCE IF NOT EXISTS public.policy_1000
INCREMENT 1
START 46656 -- 1,000
MINVALUE 46656 -- 1,000
MAXVALUE 1679615 -- Z,ZZZ
CACHE 10;
ALTER SEQUENCE public.policy_1000 OWNER TO chorke_init_pro;
COMMENT ON SEQUENCE public.policy_1000 IS 'Academia Policy Sequence';
SELECT CURRVAL('policy_1000');
SELECT SETVAL('policy_1000', 0);
ALTER SEQUENCE IF EXISTS policy_1000 RESTART 1;
SELECT NEXTVAL('policy_1000')::int AS policy_id;
Execute DDL
Anonymous Blocks
DO
$BODY$
DECLARE
fiscal_year_ukc character varying;
fiscal_drop_ukc character varying := 'ALTER TABLE fiscal_year';
BEGIN
SELECT
INTO fiscal_year_ukc con.conname
FROM pg_catalog.pg_constraint con
INNER JOIN pg_catalog.pg_class rel ON rel.oid = con.conrelid
INNER JOIN pg_catalog.pg_namespace nsp ON nsp.oid = connamespace
WHERE
nsp.nspname = 'public' AND
rel.relname = 'fiscal_year' AND
con.conname LIKE 'uk_%';
IF fiscal_year_ukc IS NOT NULL THEN
fiscal_drop_ukc := fiscal_drop_ukc || ' '
'DROP CONSTRAINT ' || fiscal_year_ukc;
EXECUTE fiscal_drop_ukc;
END IF;
END; $BODY$;
Create Sequence
CREATE OR REPLACE FUNCTION fn_get_policy_no(IN agent_no character varying)
RETURNS integer
LANGUAGE 'plpgsql'
AS $BODY$
DECLARE
policy_no integer;
sequence_name character varying := 'policy_' || agent_no;
create_sequence character varying := 'CREATE SEQUENCE IF NOT EXISTS ' || sequence_name;
BEGIN
create_sequence := create_sequence || ' '
'INCREMENT 1 '
'START 46656 ' -- 1,000
'MINVALUE 46656 ' -- 1,000
'MAXVALUE 1679615 ' -- Z,ZZZ
'CACHE 10 ';
EXECUTE create_sequence;
SELECT INTO policy_no NEXTVAL(sequence_name)::int;
RETURN policy_no;
END;
$BODY$;
SELECT fn_get_policy_no('0001');
Create Sequence with Comment
CREATE OR REPLACE FUNCTION fn_get_policy_no(IN agent_no character varying, IN agent_desc character varying)
RETURNS integer
LANGUAGE 'plpgsql'
AS $BODY$
DECLARE
policy_no integer;
sequence_name character varying := 'policy_' || agent_no;
create_sequence character varying := 'CREATE SEQUENCE IF NOT EXISTS ' || sequence_name;
comment_sequence character varying := 'COMMENT ON SEQUENCE ' || sequence_name ||' IS ''' || agent_desc || '''';
BEGIN
create_sequence := create_sequence || ' '
'INCREMENT 1 '
'START 46656 ' -- 1,000
'MINVALUE 46656 ' -- 1,000
'MAXVALUE 1679615 ' -- Z,ZZZ
'CACHE 10 ';
EXECUTE create_sequence;
EXECUTE comment_sequence;
SELECT INTO policy_no NEXTVAL(sequence_name)::int;
RETURN policy_no;
END;
$BODY$;
SELECT fn_get_policy_no('0001', 'Fareast Islami Life Insurance');
-- @Query(value="SELECT fn_get_policy_no(:agent_no, :agent_desc)", nativeQuery = true)
-- Integer getPolicyNo(@Param("agent_no") String agentNo, @Param("agent_desc") String agentDesc);
Create From Table Exclude Column
DO
$BODY$
DECLARE
create_table character varying;
BEGIN
SELECT INTO create_table 'CREATE TABLE IF NOT EXISTS invoice__temp AS SELECT ' || STRING_AGG(isc.column_name, ', ') || ' FROM invoice'
FROM (SELECT * FROM information_schema.columns WHERE table_name = 'invoice' AND
column_name NOT IN ('id', 'created_on') ORDER BY ordinal_position ASC) isc;
EXECUTE create_table;
END; $BODY$;
Date & Time
SELECT EXTRACT(CENTURY FROM TIMESTAMP '2013-10-10 10:10:10');
SELECT DATE_TRUNC('year', TIMESTAMP '2013-10-10 10:10:10');
SELECT DATE_TRUNC('hour', TIMESTAMP '2013-10-10 10:10:10');
SELECT AGE(TIMESTAMP '2023-10-10', TIMESTAMP '1983-10-10');
SELECT EXTRACT(DAY FROM TIMESTAMP '2013-10-10 10:10:10');
SELECT DATE_PART('day', TIMESTAMP '2013-10-10 10:10:10');
SELECT DATE_PART('hour', INTERVAL '4 hours 3 minutes');
SELECT JUSTIFY_INTERVAL(INTERVAL '1 mon -1 hour');
SELECT JUSTIFY_HOURS(INTERVAL '27 hours');
SELECT JUSTIFY_DAYS(INTERVAL '35 days');
SELECT AGE(TIMESTAMP '1983-10-10');
SELECT CURRENT_TIMESTAMP(2)
SELECT CURRENT_TIMESTAMP;
SELECT LOCALTIMESTAMP;
SELECT CURRENT_TIME;
SELECT CURRENT_DATE;
Restoring the Dump
export dumps pg_dump academia > academia_20200210_1020.sql pg_dump academia| gzip > academia_2020021.sql.gz
SELECT pg_terminate_backend(pid)
FROM pg_stat_activity WHERE datname IN
(
'academia'
);
DROP DATABASE IF EXISTS academia;
CREATE DATABASE academia;
restore dumps psql academia < academia_20200210_1020.sql gunzip -c academia_20200210_1020.sql.gz | psql academia
OS User Auth
CREATE DATABASE academia;
CREATE USER academia WITH LOGIN;
GRANT ALL PRIVILEGES ON DATABASE academia TO academia;
-- CREATE USER academia WITH ENCRYPTED PASSWORD 'sadaqah!';
echo "os-map academia academia" >> $PG_DATA/pg_ident.conf
echo "host all all 127.0.0.1/32 ident map=os-map" >> $PG_DATA/pg_hba.conf
pg_ctl --pgdata="$PG_DATA" reload;
DBLink Exec
DO $BODY$
DECLARE
p_connstr text := 'hostaddr=127.0.0.1 port=5432 dbname=academia';
BEGIN
PERFORM dblink_exec(p_connstr, format('UPDATE commission SET type = ''%s'' WHERE type = ''%s''', 'new_type', 'old_type'));
END;
$BODY$;
Encode/Decode
SELECT decode('Chorke Academia, Inc.', 'escape'); -- text to binary
SELECT encode('Chorke Academia, Inc.'::bytea, 'escape'); -- binary to text
SELECT convert_to('Chorke Academia, Inc.', 'UTF8'); -- text to binary
SELECT convert_from('Chorke Academia, Inc.'::bytea, 'UTF8'); -- binary to text
SELECT encode('Chorke Academia, Inc.'::bytea, 'base64'); -- text to base64
SELECT convert_from(decode('Q2hvcmtlIEFjYWRlbWlhLCBJbmMu', 'base64'), 'UTF8'); -- base64 to text
Large Object
\dfS lo_*
-- lo_open(lobjId oid, mode integer) returns integer
-- The mode parameter to lo_open uses two constants:
-- INV_READ = 0x20000
-- INV_WRITE = 0x40000
-- INV_READ_WRITE = 0x60000
SELECT x'20000'::int; -- SELECT cast(x'20000' AS integer); -- 131072
SELECT x'40000'::int; -- SELECT cast(x'40000' AS integer); -- 262144
SELECT x'60000'::int; -- SELECT cast(x'60000' AS integer); -- 393216
SELECT (x'20000'|x'40000')::int -- SELECT cast(x'20000'|x'40000' AS integer); -- 393216
-- example: a
SELECT lo_from_bytea(0, lo_get(photo::oid)) FROM employee;
SELECT lo_get(photo::oid), lo_get(resume::oid) FROM employee;
SELECT encode(lo_get(photo::oid), 'escape'), encode(lo_get(resume::oid), 'escape') FROM employee; -- binary to text
SELECT encode(lo_get(photo::oid), 'base64'), encode(lo_get(resume::oid), 'base64') FROM employee; -- binary to base64
-- example: b
SELECT resume FROM profile; -- oid
SELECT lo_get(resume::oid) FROM profile; -- binary
SELECT encode(lo_get(resume::oid), 'escape') FROM profile; -- text
SELECT encode(lo_get(resume::oid), 'base64') FROM profile; -- base64
SELECT convert_from(lo_get(resume::oid), 'UTF8') FROM profile;
SELECT convert_from(loread(lo_open(resume::oid, x'40000'::int), x'40000'::int), 'UTF8') FROM profile;
SELECT lowrite(lo_open(resume::oid, x'60000'::int), convert_to('Chorke Academia, Inc.', 'UTF8')) FROM profile;
Knowledge
comment and uncomment sql sed -i -e 's|^|-- |g' src/main/resources/db/migration/*.sql sed -i -e 's|^-- ||g' src/main/resources/db/migration/*.sql
-- psql
CREATE DATABASE chorke_init_pro;
CREATE USER chorke_init_pro WITH ENCRYPTED PASSWORD 'pa55w0rd';
GRANT ALL PRIVILEGES ON DATABASE chorke_init_pro TO chorke_init_pro;
-- \q
-- revoke public connect from specific database after creation
REVOKE CONNECT ON DATABASE academia_ebis_dev FROM PUBLIC;
-- revoke public connect for all future databases
REVOKE CONNECT ON DATABASE template1 FROM PUBLIC;
SELECT
con.conname,
nsp.nspname,
rel.relname
FROM pg_catalog.pg_constraint con
INNER JOIN pg_catalog.pg_class rel ON rel.oid = con.conrelid
INNER JOIN pg_catalog.pg_namespace nsp ON nsp.oid = connamespace
WHERE
nsp.nspname = 'public' AND
rel.relname = 'fiscal_year' AND
con.conname LIKE 'uk_%';
ALTER USER academia WITH SUPERUSER;