Rounding a number(from advice given here)

[Hal Rosser]

From previous postings, I assume that the round method of the Math object
leaves too much to chance, and should not be used. It has also been
suggested that rounding be done by converting the number to string - then
use string manipulation.
So - with that in mind, please check the validity of the following code:
// the args: n is the number to round, and x the number of places
function roundNtoXplaces(n, x){
// move the decimal over x places
n = n * (Math.pow(10, x));
// coerce the number to a string
n = "" + n;
// split the string at the new decimal place
var numParts = n.split(".");
// if the decimal part would come after "5" if sorted alphabetically,
// then add 1 to the int part
if (numParts[1]+"" >= "5"){
numParts[0]++;
}
//coerce the number back to a string
// This string contains the numbers we will be returning
numParts[0] += "";
//get the length of the string
var len = numParts[0].length;
// since we moved the decimal over to work with it, now we put it back
// intPart is the part before the decimal
var intPart = numParts[0].substr(0, (len - x));
//decimalPart is the part of the string after the decimal
var decimalPart = numParts[0].substr( len - x, x );
var outString =intPart + "." + decimalPart;
return outString;

 

[Hal Rosser]

return outString;
I left off the closing "}" curly-braces when pasting the code.
And of course I know the number of lines can be condensed
and the code could be obfusicated beyond recognition.
The extra steps and comments are there for clarity.

 

[Dr John Stockton]

JRS: In article <[email protected]>, dated
Fri, 7 Apr 2006 14:12:11 remote, seen in Hal

From previous postings, I assume that the round method of the Math object
leaves too much to chance, and should not be used. It has also been
suggested that rounding be done by converting the number to string - then
use string manipulation.

To show potential intelligence, you could have refrained from starting a
new thread and you could have indented your code, as is customary, to
show structure.

You could also have tested it.

So - with that in mind, please check the validity of the following code:
// the args: n is the number to round, and x the number of places
function roundNtoXplaces(n, x){
// move the decimal over x places
n = n * (Math.pow(10, x));
// coerce the number to a string
n = "" + n;
// split the string at the new decimal place
var numParts = n.split(".");
// if the decimal part would come after "5" if sorted alphabetically,
// then add 1 to the int part
if (numParts[1]+"" >= "5"){
numParts[0]++;
}

To test whether the first character of a string is 5 or more, knowing
that it will be in 0..9, one might use S.charAt()>="5" - it seems OK
with an empty string - undertested.

//coerce the number back to a string
// This string contains the numbers we will be returning
numParts[0] += "";
//get the length of the string
var len = numParts[0].length;
// since we moved the decimal over to work with it, now we put it back
// intPart is the part before the decimal
var intPart = numParts[0].substr(0, (len - x));
//decimalPart is the part of the string after the decimal
var decimalPart = numParts[0].substr( len - x, x );
var outString =intPart + "." + decimalPart;
return outString;

Missing } shows carelessness.

roundNtoXplaces(0.0, 2) .1
roundNtoXplaces(0.07, 2) .7
roundNtoXplaces(90.07, 2) 90.08
roundNtoXplaces(-90.07, 2) -90.06
roundNtoXplaces(0.0999, 2) .10 **
roundNtoXplaces(11.34, 5) 11.34001

** Numerically correct but format deprecated.

 

[Hal Rosser]

To show potential intelligence, you could have refrained from starting a

new thread and you could have indented your code, as is customary, to
show structure.

*** OK ** but potential intelligence may be too fleeting for me

You could also have tested it.

*** I did, but you did a much better job at that.

To test whether the first character of a string is 5 or more, knowing
that it will be in 0..9, one might use S.charAt()>="5" - it seems OK
with an empty string - undertested.

*** I wanted to see if the string, if sorted alphabetically, would come on
or after a "5.

Missing } shows carelessness.

*** yeah, When I copied it from the web page I tested it on, the "}" didn't
make the copy buffer. I inadvertantly left it off.
****
Thanks for taking the time to test the code.
The bad output examples you pointed out helped a lot in debugging.
The revised code fixes those errors, and I also tested the function with
ramdom numbers and random 'places', and seems to be working now. (I wrote a
function to test it 100 times and it looks ok.)
(I don't know what it was about ".07", but adding zeroes to the end of the
string fixed that part. I think an empty string must be rounding to 1.)
**
***** Here's the revised code **
// the args: n is the number to round, and x the number of places
// This function formats to x decimal places - padding with trailing
// zeroes if required
function roundNtoXplaces(n, x){
var itWasNegative = false;
if (n < 0){
n*=-1;
itWasNegative = true;
}
// move the decimal over x places
n = n * (Math.pow(10, x));
// coerce the number to a string
n = "0" + n;
//** to assure there is something after the deci9mal to work with
if (n.indexOf(".") < 0){
n += ".0000"
}
// split the string at the new decimal place
var numParts = n.split(".");
//** to avoid empty array, add zeroes to the end of the string***
numParts[1] += "" + "00000";
// if the decimal part would come after "5" if sorted
// then add 1 to the int part
if (numParts[1]+"" >= "5"){
numParts[0]++;
}
// coerce the number back to a string
// This string contains the numbers we will be returning
numParts[0] = "" + numParts[0];
//get the length of the string
var len = numParts[0].length;
// since we moved the decimal over , now we put itback
// intPart is the part before the decimal
var intPart = numParts[0].substr(0, (len - x));
intPart *=1; // removes leading zeroes
//decimalPart is the part of the string after the decimal
var decimalPart = numParts[0].substr( len - x, x );
if (itWasNegative){
intPart *=-1;
}
var outString = intPart + "." + decimalPart;
return outString;
}// Here's that pesky closing curly-brace ;-)

 

[Dr John Stockton]

JRS: In article <[email protected]>, dated Fri, 7
Apr 2006 21:36:55 remote, seen in Hal Rosser

*** OK ** but potential intelligence may be too fleeting for me

Duplicating News in Mail is quite unnecessary. Doing do without
explicit wanting in both media is extremely bad manners.

 

[Hal Rosser]

JRS: In article <[email protected]>, dated Fri, 7
Apr 2006 21:36:55 remote, seen in Hal Rosser


Duplicating News in Mail is quite unnecessary. Doing do without
explicit wanting in both media is extremely bad manners.

My apologies. Calling folks stupid is ok, but not an inadvertant mouse
click?
I carelessly but unintentionally clicked "reply" instead of "reply group".
The message was returned undeliverable, anyway.
Get back to work.

 

[optimistx]

JRS: In article <[email protected]>, dated Fri, 7
Apr 2006 21:36:55 remote, seen in Hal Rosser

Duplicating News in Mail is quite unnecessary. Doing do without
explicit wanting in both media is extremely bad manners.

Thank you for teaching gad manners for us.

: