Question

I have an working query and not sure if its efficient or not.

Table columns:

1. p_type: can be STRING sell, cancelsell, bank, cancelbank
2. id
3. fid
4. unit

Operation to be performed:

1. SUM(unit) for each process type sell,cancelsell,bank,cancelbank.
2. SUM(UNIT for sell) – SUM(unit for cancelsell) , SUM(unit for bank) – SUM(unit for cancelbank)

NOTE: I am also checking for id and fid in each select though I need to check with same values for all select.

The existing query:

    select sell-cancelsell scount, bank-cancelbank bcount from

    (select sum(a.unit) as sell from table1 a where   
    a.p_type = 'Sell' and a.id=1 and a.fid=2 ),

    (select sum(c.unit) as cancelsell from table1  c where      
    c.p_type = 'CancelSell' and c.id=1 and c.fid=2),

    (select sum(b.unit) as bank from table1  b where     
    b.p_type = 'Bank' and b.id=1 and b.fid=2),

    (select sum(d.unit) as cancelbank from table1  d where  
    d.p_type = 'CancelBank' and d.id=1 and d.fid=2)

Is good enough? If anyone can suggest a way to make it more efficient that would be great.

Answer 1

You can do this

select 
sum(Case when a.p_type = 'sell' then a.unit else null end) as sellUnit, 
sum(Case when a.p_type = 'CancelSell' then a.unit else null end) as CancelSellUnit,
sum(Case when a.p_type = 'Bank' then a.unit else null end) as BankUnit ,
sum(Case when a.p_type = 'CancelBank' then a.unit else null end) as CancelBankUnit  
from table1 a where and a.id=1 and a.fid=2