Link to home
Start Free TrialLog in
Avatar of joesmow
joesmow

asked on

How to round down in VB?

What is the best way to round down a number?
Avatar of Ryan Chong
Ryan Chong
Flag of Singapore image

do while Conditon
   i = i - j
loop
ASKER CERTIFIED SOLUTION
Avatar of AzraSound
AzraSound
Flag of United States of America image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Oops..
Debug.Print Format(5459.455, "##,##0.00")

For more information place cursor on 'Format' word and press F1.

Cheers
Avatar of nilapenn
nilapenn

use the found function

for (eg)

round(45.34,2)=45
round(45.55)=46
roudn(45.55,1)=45.6
Warning Round has some strange bugs:

If you enter an immediate window and you type:
 
?round(0.5,0)

The .5 should round up to 1 but it produces a result of  0

Scratch head..............then try

?round(.95,1)

But in this case the .05 correctly rounds up.

I am not qualified to pass judgement on Microsoft's QA procedures.  The problems stem from how numerics are internally represented.

I would therefore suggest that you should not used either Round, Fix or Int. You should create your own function so that at some time in the future, if some smartypants find a problem with your rounding function you only have one place to change.

The following function will provide accurate results for financial applications.  Roundings for both negative and positive numbers are consistent. Which means that credit notes will always have the same total value as invoices.

Public Function Rounder(Value, DecimalPlaces As Integer)

Public Function Rounder(Value, DecimalPlaces As Integer)
If DecimalPlaces <= 4 Then
    Rounder = (Int(Abs(Value) * (10 ^ DecimalPlaces) + 0.50001) * (10 ^ -DecimalPlaces)) * Sgn(Value)
Else
    Rounder = (Int(Abs(Value) * (10 ^ DecimalPlaces) + (0.5 + (10 ^ (-DecimalPlaces - 2))))) * (10 ^ -DecimalPlaces) * Sgn(Value)
End If
End Function

Hope this helps, inthedark.
Warning Round has some strange bugs:

If you enter an immediate window and you type:
 
?round(0.5,0)

The .5 should round up to 1 but it produces a result of  0

Scratch head..............then try

?round(.95,1)

But in this case the .05 correctly rounds up.

I am not qualified to pass judgement on Microsoft's QA procedures.  The problems stem from how numerics are internally represented.

I would therefore suggest that you should not used either Round, Fix or Int. You should create your own function so that at some time in the future, if some smartypants find a problem with your rounding function you only have one place to change.

The following function will provide accurate results for financial applications.  Roundings for both negative and positive numbers are consistent. Which means that credit notes will always have the same total value as invoices.

Public Function Rounder(Value, DecimalPlaces As Integer)

Public Function Rounder(Value, DecimalPlaces As Integer)
If DecimalPlaces <= 4 Then
    Rounder = (Int(Abs(Value) * (10 ^ DecimalPlaces) + 0.50001) * (10 ^ -DecimalPlaces)) * Sgn(Value)
Else
    Rounder = (Int(Abs(Value) * (10 ^ DecimalPlaces) + (0.5 + (10 ^ (-DecimalPlaces - 2))))) * (10 ^ -DecimalPlaces) * Sgn(Value)
End If
End Function

Hope this helps, inthedark.
But inthedark, you say, "I would ... suggest ... not [using]... Int"

then your code uses "int":

   Rounder = (Int(Abs(Value) * (10 ^ DecimalPlaces) + 0.50001) * (10 ^ -DecimalPlaces)) * Sgn(Value)

Granted, there is really no other practical way to truncate a value's integer.  You could convert to string, find the decimal, then remove it and convert back to numeric, but that seems like a silly way.

--
Basically, the functions you mention have problems with the decimal portion, so one way to deal with this is to "convert" to a larger scaled number:

Instead of int(123.5), you can try int((123.5*10-9)/10) which scales it to an integer, subtracts a rounding value (always in 9s), then scales back and removes the decimal portion.  I don't know if this has the same decimal-to-binary-to-decimal conversion problems, but they may be more limited.

Personally, I think it's overkill except for the most critical of applications, and I'd simply use Int(x).
rspahitz, you are quite right I used Int. I only posted my comment as I read that somebody else had suggested using the Round function. The points that I was trying to make: keep control - always use your own function for rounding so that you can future-proof your application against changes in Microsoft's current interpretation of rounding. And that I wouldn't like a VB newbie reading this to think that rounding was a simple subject.

AzraSound has already answered the question which was "What is the best way to round down a number?"
Yes.  I think most of the other posts before mine addressed the issue, and you're quite right about the rounding issue.  And AzraSound has the simplest (and first) solution that will probably work.