sql - mysql execution plan explanation -

there following query

select count(*) my_table my_t my_t.c1 = '3' , my_t.c2 = '123' 

this table has index on (c1, c2) , both these columns have not null constraint.

its execution plan:

# id, select_type, table, type, possible_keys,    key, key_len,         ref, rows,    1,      simple,  my_t,  ref,        my_idx, my_idx,     157, const,const,    1, using where; using index 

it uses where, purpose? both columns not null index contain records , necessary count them.

am missing something?

the documentation says:

using index

the column information retrieved table using information in index tree without having additional seek read actual row. strategy can used when query uses columns part of single index.

if extra column says using where, means index being used perform lookups of key values. without using where, optimizer may reading index avoid reading data rows not using lookups. example, if index covering index query, optimizer may scan without using lookups.

as far understand above, using index means reading index specified in key column enough data needed compute result; there no need read table data. doesn't how index information used.

using where 1 says values read index mentioned in key column compared against values of columns or constants indicated in ref column of output. ref column contains const, const because values use compare against string constants ('3' , '123').