Question

Possible Duplicate:
COUNT() vs. COUNT(1) vs. COUNT(pk): which is better?
count() and count(column_name), what's the diff?
count(*) vs count(column-name) – which is more correct?

The benefit of using count(*) in a select statement is that I can use it with any table and that makes automating scripts easier:

count_sql = 'select count(*) ' + getRestOfSQL('tablename');

But, is it less efficient than using count(specific_field)?

Answer 1

For InnoDB

If specific_field is not nullable, they are equivalent and have the same performance.

If specific_field is nullable, they don’t do the same thing. COUNT(specific_field) counts the rows which have a not null value of specific_field. This requires looking at the value of specific_field for each row. COUNT(*) simply counts the number of rows and in this case can be faster as it does not require examining the value of specific_field.

For MyISAM

There is a special optimization for the following so that it does not even need to fetch all rows:

SELECT COUNT(*) FROM yourtable