count (*) is calculated on the number of lines, comprising a NULL
COUNT (column) having a number of lines for a particular column value is calculated, it does not contain a NULL value.
count () there is a way to use the results count (1) and the usage count (*) is the same.
In any case select count (*) from tablename is the best choice;
Minimizing select count (*) from tablename where COL = 'value' Such a query;
Eliminate select count (COL) tablename where COL2 = 'value' emergence from.
If the table has no primary key, then count (1) (*) faster than the count.
If there is a primary key, then count (primary key, the primary key) (*) faster than the count.
If the table has only one field, count (*) the fastest.
count (1) Like count (primary key), only scans the primary key. count (*) with the same count (non-key), scan the entire table. The former obviously faster.
If the statement is contained where, in the conditions index takes precedence where.