Question

I have a strange problem with the float data type.
Please find the screen shot below.

Query 1 gives a record. Where as Query 2 gives no record. netbk_amt is of type float in the table.
Also, when i change the 1st query to >= (convert(float, 100001)), it doesn’t give any records.

Looks like the value is > 1000000 but < 1000001. But in the table it is shown as 1000000.

Please help me to find what went wrong with this query.

Queries:

select co_id, SUM(netbk_amt) from its_deal_sum where co_id = 19237 
group by co_id
having convert(float,SUM(netbk_amt)) >= convert(float,1000000)

select co_id, SUM(netbk_amt) from its_deal_sum where co_id = 19237 
group by co_id
having convert(float,SUM(netbk_amt)) <= convert(float,1000000)

I kind of did this as suggested in comments,

select co_id, SUM(netbk_amt) from its_deal_sum where co_id = 19237 group by co_id having convert(money,SUM(netbk_amt)) = convert(money,1000000) 

as a temporary fix. But looking for a robust fix with the float, ignoring the fractions.

Answer 1

If

ignoring the fractions

is all what you want – it is very easy.

DECLARE @f float=2.000000000001
SELECT @f,FLOOR(@f)

Result set:

(No column name)    (No column name)
2.000000000001      2