Question

I have a table using innodb. I know the table has roughly 89 million rows. Using

SELECT COUNT(*) FROM table;

takes about five minutes to run. I know that innodb is not optimized for unconditional COUNT(*) queries. How can I restructure the query to give me a count faster? Would just adding WHERE 1=1 work, or do I need to query a specific field?

I know I can get the approximate number of rows using SHOW TABLE STATUS, but I want to get the value in a PHP script, and it seems like there is a lot to dig through using that method.

Answer 1

If you are OK with the estimated number and just don’t want to mess with running SHOW TABLE STATUS from PHP, you can use the information_schema DB:

SELECT TABLE_ROWS FROM information_schema.tables
WHERE TABLE_SCHEMA = 'my_db_name' 
AND TABLE_NAME = 'my_table_name';