Rounding... sort of

[Guest]

Hello everyone. I am trying to take a number and round it down. For example, I want to change 38 to 30 or 32 to 30 or 59 to 50. Anyone know how to do that?

[-__cfSearching__-]

Subtract modulus 10 from the number.

ie
38 - (38 MOD 10) ...OR
38 - (8) = 30

[Oğuz Demirkapı]

The function to find the closest integer is "Int". But depending on your requirements you can figure out different techniques.

[-__cfSearching__-]

I think they want to round down to the nearest multiple of 10. In which case int would not suffice.

[Guest]

That is correct. I did find a temporary solution to this but it's not clean and I know there's something better.

[-__cfSearching__-]

Did you try using modulus? It seems to do what you are looking for.

newNumber = origNumber - (origNumber MOD 10) .

[Guest]

No because I figured it out prior to that reply. However if I decide to not be lazy I will try MOD.

[BKBK]

Idesdema wrote:
I am trying to take a number and round it down. For example, I want to change 38 to 30 or 32 to 30 or 59 to 50.

Sounds like rounding down to the nearest ten. Then there is one function tailor-made for it: fix().

<cfset x=38>
<cfoutput>#fix(x/10)*10#</cfoutput>

----------------- you may ignore what follows --------------

Similarly, to round down to nearest dozen:

<cfset x=38>
<cfoutput>#fix(x/12)*12#</cfoutput>

To round down to nearest score:

<cfset x=38>
<cfoutput>#fix(x/20)*20#</cfoutput>

and so on.

[Newsgroup_User]

BKBK wrote:
> Idesdema wrote:
> Sounds like rounding down to the nearest ten. Then there is one function
> tailor-made for it: fix().

Why fix() over int()? I was about to suggest the old school int()
solution which is basically exactly the same as your fix() version.

<cfoutput>#int(38/10)*10#</cfoutput>

> <cfset x=38>
> <cfoutput>#fix(x/10)*10#</cfoutput>

> Similarly, to round down to nearest dozen:
>
> <cfset x=38>
> <cfoutput>#fix(x/12)*12#</cfoutput>
<cfoutput>#int(x/12)*12#</cfoutput>
>
> To round down to nearest score:
>
> <cfset x=38>
> <cfoutput>#fix(x/20)*20#</cfoutput>
<cfoutput>#int(x/20)*20#</cfoutput>

[-__cfSearching__-]

I think it depends on the desired results when the number is negative. Or if negative values are even a factor here.

<cfset mult = 10>
<cfset x = 38>
<cfoutput>
1. fix = #(fix(x/mult)*mult)#<br>
1. int = #(int(x/mult)*mult)#<br>
1. mod = #(x - (x mod mult))#<hr>
</cfoutput>

<cfset mult = 10>
<cfset x = -38>
<cfoutput>
2. fix = #(fix(x/mult)*mult)#<br>
2. int = #(int(x/mult)*mult)#<br>
2. mod = #(x - (x mod mult))#<br>
</cfoutput>

Results:
1. fix = 30
1. int = 30
1. mod = 30
--------------------------------------------------------------------------------
2. fix = -30
2. int = -40
2. mod = -30

[BKBK]

Ian Skinner wrote:
Why fix() over int()?

Just personal preference. After all fix() and int() are basically equivalent. Mathematically. The point of my post is the trick to divide and multiply by 10.

[Newsgroup_User]

BKBK wrote:
> Ian Skinner wrote:
> Why fix() over int()?
>
> Just personal preference. After all fix() and int() are basically equivalent. Mathematically.
>
>

Basically but not quite the same. Which can lead one into a lot of
trouble if one does assume they are in fact the same.

[BKBK]

Ian Skinner wrote about fix() and int():
not quite the same

True. However, I didn't say they are the same. I said they are mathematically equivalent.

In all the programming languages I know, there are relationships between fix() and int() that ensure that, whatever you can do with fix(), you can also do with int(). For example, I found these relationships in CFML

fix(x) = sgn(x) * int(abs(x))
int(x) = sgn(x)*fix(abs(x)) + sgn(sgn(x-fix(x))+1) - 1