Solved

# Why do these yield different results ?

Posted on 2009-05-18
377 Views
As you can see by the code the latter query has more decimals why cant I get the first query to yield the same result ?
``````SELECT A,B,A/B FROM

(

SELECT colA, colB FROM thetable

) C

SELECT A,B,A/B FROM

(

SELECT		1.00000000000000000000 AS A, 1.40673489600528227000 AS B

) C

And the Result

1.00000000000000000000	1.40673489600528227000	0.710865

1.00000000000000000000	1.40673489600528227000	0.71086599389814597
``````
0
Question by:Pall Palsson
• 4
• 3
• 2
• +1

LVL 29

Expert Comment

ID: 24418247
my *guess* is that it has something to do with the datatypes in thetable.
I can't explain why from your results but:
select 1.00000000000000000000/convert(float,1.40673489600528227000) = 0.710865993898146
and
select 1.00000000000000000000/convert(real,1.40673489600528227000) = 0.710866
0

LVL 31

Expert Comment

ID: 24418349
it is completely depends on the datatype you use in your table or in your calculation. have a look at following queries and its results printed after query.

``````SELECT A,B,A/B FROM

(

SELECT             1.00000000000000000000 AS A, 1.40673489600528227000 AS B

) C --0.71086599389814597

SELECT A,B,A/B FROM

(

SELECT             1.00000000000000000000 AS A, 1.40 AS B

) C --0.714285714285714285714285

SELECT A,B,A/B FROM

(

SELECT             1 AS A, 1.40 AS B

) C --0.714285
``````
0

LVL 4

Expert Comment

ID: 24418494
Use the format below to make sure division occurs properly and to control decimal output of the result:
SELECT convert(decimal(18,2), (convert(decimal(18,2), A)  /  convert(decimal(18,2), D) ) )

See http://www.experts-exchange.com/Programming/Languages/SQL_Syntax/Q_24416319.html
0

Author Comment

ID: 24420731
Hi, both fields (ColA and ColB) are decimal(38, 20)

If you look at the results  the upper line is the result from the Columns and the lower from just typing in the values in the select.

If I use + or - i.e. A+B or A-B I get the correct amount of decimals.  If I try to use Cast or Convert to decimal(38, 20) I Get the same result 0.710865 (0.71086500000000000)

If I multiply or divide the cutoff happens.

So neither of your proposals work.
``````1.00000000000000000000  1.40673489600528227000  0.710865

1.00000000000000000000  1.40673489600528227000  0.71086599389814597
``````
0

LVL 4

Expert Comment

ID: 24420902
This should result in the same number of decimal places...

SELECT A,B,convert(decimal(38,20), A/B)  FROM
(
SELECT colA, colB FROM thetable
) C

SELECT A,B,convert(decimal(38,20), A/B)  FROM
(
SELECT            1.00000000000000000000 AS A, 1.40673489600528227000 AS B
) C
0

Author Comment

ID: 24421057
Tim:

What you propose results in 0.71086500000000000 which is not the same as 0.71086599389814597

Thus it doesnt work
0

LVL 31

Expert Comment

ID: 24421070

SELECT A,B,convert(decimal(38,20), A)/convert(decimal(38,20), B)  FROM
(
SELECT colA, colB FROM thetable
) C

0

LVL 4

Expert Comment

ID: 24421550
The code below tells the equation to limit B to 2 decimal places and should create the same result.   You actually need to do this for A as well.  The only way you can get the same result is to use the same number of decimal places for B each time and for A each time. This is back to my original response.

SELECT A,convert(decimal(38,2), B),convert(decimal(38,20), A/B)  FROM
(
SELECT colA, colB FROM thetable
) C

SELECT A,convert(decimal(38,2), B),convert(decimal(38,20), A/B)  FROM
(
SELECT            1.00000000000000000000 AS A, 1.40673489600528227000 AS B
) C
0

Author Comment

ID: 24423762
Hi,

I have tried to to use convert(decimal(38,20) AND cast like this X() represents what i tried to convert/cast

X(A)/B
X(A)/X(B)
A/X(B)
X(X(A)/X(B))
X(A/X(B))
X(X(A)/B)

with no success.

However QPR's solution of using float comes closest to yielding the correct result.
``````cast/convert   result

correct:       0.71086599389814597

float:         0.71086599389814600

decimal:       0.71086500000000000

noone:         0.710865
``````
0

LVL 4

Accepted Solution

TimSledge earned 500 total points
ID: 24423868
I just realized you need to use  Convert(Decimal(32,1) to get the same results.  When you use anything higher than 1 in the 2nd position (controls how many places to the right of the decimal ) you are not multiplying by the same numbers.

You can never get the same answer when multiplying by 1.4 one time then 1.40673489600528227000 another time.

Convert(Decimal(TotalDigits,DigitsToRightOfDecimal) ItemToConvert) is a way to make the two multipliers the same.  Convert(Decimal(32,1), 1.4) and Convert(Decimal(32,1), 1.40673) will both multiply by 1.4.

If you go one step higher (32,2), you will be multiplying by 1.40 and 1.41 (1.406 rounded to two places).

Try the earlier suggestion, but use 32,1 instead of 32,2.
0

## Featured Post

by Mark Wills Attending one of Rob Farley's seminars the other day, I heard the phrase "The Accidental DBA" and fell in love with it. It got me thinking about the plight of the newcomer to SQL Server...  So if you are the accidental DBA, or, simpâ€¦
How to leverage one TLS certificate to encrypt Microsoft SQL traffic and Remote Desktop Services, versus creating multiple tickets for the same server.
This demo shows you how to set up the containerized NetScaler CPX with NetScaler Management and Analytics System in a non-routable Mesos/Marathon environment for use with Micro-Services applications.
This tutorial demonstrates a quick way of adding group price to multiple Magento products.