About the difference between count (1), count (*), and count (column name)

count(1) and count(*) There is no difference .

count( Column ) and count(*) It's actually as fast , If the index column is not empty ,count(*) Available to index , Just as fast at this time .

summary ： But is the conclusion really like this . It's not . In fact, in the database count(*) and count( Column ) It is simply not equivalent ,count(*) It is aimed at the whole table , and count( Column ) It is for a certain column , If this column value is empty ,count( Column ) I won't count this line . So there is no comparability between the two , Performance comparison should first consider writing equivalence , These two statements are not equivalent at all . It will lose the significance of comparison ！！！

If the query table has only one field , be COUNT(*) The fastest ;

If the query table has no primary key , be COUNT(1) Than COUNT(*) fast ;

If the query table has a primary key , be COUNT( Primary key ) Than COUNT(*) fast ,count( Primary key column )》count(1)》count(*);

COUNT(*) Return the total number of rows , contain NULL value ; COUNT( Name ) Return the total number of rows , Not included NULL value ; DISTINCT COUNT(*) Return the total number of rows , Remove duplicates without NULL value