I'm doing a statistic job like this:

SUM    |COND1 |COND2 |...
--------------------------
100    |80    | 70   |...

The SUM result is calculated from multiple tables, and the CONDs are subset of that.

I wrote a sql like this:

select tmp1.id,sum,cond1,cond2 as count from (
   select id, count(*) as sum from table_1 
   group by table1.id) tmp1
 left join ( 
   select id, count(*) as cond1 from table1
   where condition1
   group by table1.id) tmp2 on tmp1.id=tmp2.id
 left join ( 
   select id, count(*) as cond2 from table1
   where condition2
   group by table1.id) tmp3 on tmp1.id=tmp3.id

the problem is that this is very poor efficiency, it will be better if i could use the result of tmp1, but i don't know how to do that.

update: simplified the sql, what i mean is how to reuse the first select result of table_1