select count(*) 的原理是什么？

wdjk1999

最初由 rollingpig 发布
[B]count(*) 就是类似于
declare
cnt number default 0;
begin
for i in (select  rowid from table_name) loop
cnt :=cnt + 1;
end loop;
end;

而 count(1) 就类似于

declare
cnt number default 0;
begin
for i in (select  1 as selected from table_name) loop
if(selected is not null) then
cnt :=cnt + 1;
end if;
end loop;
end;

而 count(col_name) 就类似于

declare
cnt number default 0;
begin
for i in (select  col_name from table_name) loop
if(col_name is not null) then
cnt :=cnt + 1;
end if;
end loop;
end; [/B]

rolling,那你觉的三个COUNT写法那种效率更高啊

bluemoon0083

个人感觉select count(rowid) from table_name效率最高

zhaolcq

我也觉得rowid 估计更高效