MySQL dropping all indexes from table

aviv picture aviv · Sep 26, 2010 · Viewed 39k times · Source

I have a MySQL database that runs for some time now with many changes on it. Lately I looked over it and I noticed that in some cases I have doubled the index on the same field. Some Indexes are missing, and in general there is a huge mess in all the indexes.

I wants to drop all indexes from a table. Later on I have a prepared script that will run ALTER TABLE and add the relevant indexes.

Is there a way to drop all indexes from a table?

Answer

Vojtech Kurka picture Vojtech Kurka · May 17, 2016

Simple script:

-- list all non-unique indexes
SELECT table_name AS `Table`,
       index_name AS `Index`,
       GROUP_CONCAT(column_name ORDER BY seq_in_index) AS `Columns`
FROM information_schema.statistics
WHERE NON_UNIQUE = 1 AND table_schema = 'mydatabase' AND table_name = 'mytable'
GROUP BY 1,2;

-- drop all non-unique indexes
SET SESSION group_concat_max_len=10240;

SELECT CONCAT('ALTER TABLE ', `Table`, ' DROP INDEX ', GROUP_CONCAT(`Index` SEPARATOR ', DROP INDEX '),';' )
FROM (
SELECT table_name AS `Table`,
       index_name AS `Index`
FROM information_schema.statistics
WHERE NON_UNIQUE = 1 AND table_schema = 'mydatabase' AND table_name = 'mytable'
GROUP BY `Table`, `Index`) AS tmp
GROUP BY `Table`;

-- add all non-unique indexes , WITHOUT index length spec
SET SESSION group_concat_max_len=10240;
SELECT CONCAT('ALTER TABLE ', `Table`, ' ADD INDEX ', GROUP_CONCAT(CONCAT(`Index`, '(', `Columns`, ')') SEPARATOR ',\n ADD INDEX ') )
FROM (
SELECT table_name AS `Table`,
       index_name AS `Index`,
        GROUP_CONCAT(column_name ORDER BY seq_in_index) AS `Columns`
FROM information_schema.statistics
WHERE NON_UNIQUE = 1 AND table_schema = 'mydatabase' AND table_name = 'mytable'
GROUP BY `Table`, `Index`) AS tmp
GROUP BY `Table`;