drop all tables sharing the same prefix in postgres
Suppose the prefix is 'sales_'
Step 1: Get all the table names with that prefix
SELECT table_name
FROM INFORMATION_SCHEMA.TABLES
WHERE TABLE_NAME LIKE 'sales_%';
Step 2: Click the "Download as CSV" button.
Step 3: Open the file in an editor and replace "sales_ with ,sales and " with a space
Step 4: DROP TABLE sales_regist, sales_name, sales_info, sales_somthing;
To do this in one command you need dynamic SQL with EXECUTE
in a DO
statement (or function):
DO
$do$
DECLARE
_tbl text;
BEGIN
FOR _tbl IN
SELECT quote_ident(table_schema) || '.'
|| quote_ident(table_name) -- escape identifier and schema-qualify!
FROM information_schema.tables
WHERE table_name LIKE 'prefix' || '%' -- your table name prefix
AND table_schema NOT LIKE 'pg\_%' -- exclude system schemas
LOOP
RAISE NOTICE '%',
-- EXECUTE
'DROP TABLE ' || _tbl; -- see below
END LOOP;
END
$do$;
This includes tables from all schemas the current user has access to. I excluded system schemas for safety.
If you do not escape identifiers properly the code fails for any non-standard identifier that requires double-quoting.
Plus, you run the risk of allowing SQL injection. All user input must be sanitized in dynamic code - that includes identifiers potentially provided by users.
Potentially hazardous! All those tables are dropped for good. I built in a safety. Inspect the generated statements before you actually execute: comment RAISE
and uncomment the EXECUTE
.
If any other objects (like views etc.) depend on a table you get an informative error message instead, which cancels the whole transaction. If you are confident that all dependents can die, too, append CASCADE
:
'DROP TABLE ' || _tbl || ' CASCADE;
Closely related:
- Update column in multiple tables
- Changing all zeros (if any) across all columns (in a table) to... say 1
Alternatively you could build on the catalog table pg_class
, which also provides the oid
of the table and is faster:
...
FOR _tbl IN
SELECT c.oid::regclass::text -- escape identifier and schema-qualify!
FROM pg_catalog.pg_class c
JOIN pg_catalog.pg_namespace n ON n.oid = c.relnamespace
WHERE n.nspname NOT LIKE 'pg\_%' -- exclude system schemas
AND c.relname LIKE 'prefix' || '%' -- your table name prefix
AND c.relkind = 'r' -- only tables
...
System catalog or information schema?
- How to check if a table exists in a given schema
How does c.oid::regclass
defend against SQL injection?
- Table name as a PostgreSQL function parameter
Or do it all in a single DROP
command. Should be a bit more efficient:
DO
$do$
BEGIN
RAISE NOTICE '%', (
-- EXECUTE (
SELECT 'DROP TABLE ' || string_agg(format('%I.%I', schemaname, tablename), ', ')
-- || ' CASCADE' -- optional
FROM pg_catalog.pg_tables t
WHERE schemaname NOT LIKE 'pg\_%' -- exclude system schemas
AND tablename LIKE 'prefix' || '%' -- your table name prefix
);
END
$do$;
Related:
- Is there a postgres command to list/drop all materialized views?
Using the conveniently fitting system catalog pg_tables
in the last example. And format()
for convenience. See:
- How to check if a table exists in a given schema
- Table name as a PostgreSQL function parameter