Let’s assume blabla table has millions of rows, look this query;
SELECT * FROM blabla WHERE ends_at <= 1506668369 and is_processed = 0
If you have a 2 column index AS ABOVE, ‘is_processed‘ at first place and ‘ends_at‘ at second place in the index sequence, then it takes 0.0004 seconds in my server to retrieve rows.
If you have a 2 column index AS ABOVE, ‘ends_at‘ at first place and ‘is_processed‘ at second place in the index sequence, then it takes 1.0733 seconds in my server to retrieve rows.
That’s a huge difference! Think your tables as a book always and your indexes as the table of contents page when you create your indexes!
SELECT * FROM blabla WHERE ends_at <= 1506668369 and is_processed = 0
In this query, you must clearly go to is_processed=0 pages first, in those pages then you can find ends_at <= 1506668369 rows. That way will be more easy to find those rows. That’s the way of MySQL indexes works.
On the other hand, there is always a left-to-right rule in MySQL indexes.
For example; if you have a triple column index with A,B,C sequence,
Those where clauses will benefit from your index;
where A , B , C where A, B where B
However, those where clauses will not use your index;
where B,C where A, C where B where C
