Math Calculation - float numbers

Question

I am doing the following

float years = (1/31536000) * 883102.00;

and I get years = 0000000

while the actual answer is 0.0.28

Any suggestion on what might be going wrong ?

try to add the `f` suffix on each and every number – user2485710 Aug 07 '13 at 16:32 — user2485710, Aug 07 '13 at 16:32
@user2485710 `1f` is not valid C++. – Pascal Cuoq Aug 07 '13 at 16:36 — Pascal Cuoq, Aug 07 '13 at 16:36
@PascalCuoq ok I mean `1.0f`, add `.0f` to each number. – user2485710 Aug 07 '13 at 16:46 — user2485710, Aug 07 '13 at 16:46

score 3 · Accepted Answer · answered Aug 07 '13 at 16:33

3

You should instead do

float years = (1.0/31536000.0) * 883102.00;

or

float years = (1.0/31536000) * 883102.00;

might work as well. Your very first number is treated as integer.

answered Aug 07 '13 at 16:33

radarhead

568
2
13

Did the trick -Thanks – Rajeshwar Aug 07 '13 at 16:34
I personally prefer `(1.f / 31536000)` – Vassilis Nov 18 '18 at 21:42

score 3 · Answer 2 · answered Aug 07 '13 at 16:36

3

Just do

float years = 883102.00/31536000;

This will save computation. Since you are diving 1 by something and then multiplying. Alternatively just set 1 to 1.00

answered Aug 07 '13 at 16:36

Quillion

6,057
10
61
90

score 1 · Answer 3 · answered Aug 07 '13 at 16:33

1

1 is an int. The compiler then assumes that you are interested in working in ints and then 1/3153600 becomes 0. Just add a .0 to the 1 and your calc should work.

float years = (1.0/31536000) * 883102.00;

answered Aug 07 '13 at 16:33

aepryus

4,517
5
27
39

score 1 · Answer 4 · answered Aug 07 '13 at 16:33

1

That first term is being cast as an int, and thus is being rounded to 0. Try this:

float years = (1.00 / 31536000.00) * 883102.00

answered Aug 07 '13 at 16:33

gr3co

863
1
7
15

1

Why only two zeroes after the decimal point? Shouldn’t you add a few more to be safe? – Eric Postpischil Aug 07 '13 at 16:36
There is no cast here. The first term has type `int` and the second term has type `int`, so their quotient has type `int`. The value is **truncated**, not rounded. – Pete Becker Aug 07 '13 at 16:36
@PeteBecker: “Rounding” is used to mean modification of a result to fit within a destination format (e.g., IEEE 754-2008 4.3), and various modes of rounding can be used, such as rounding toward zero. – Eric Postpischil Aug 07 '13 at 16:43
@EricPostpischil - the answer doesn't "rounds toward 0", it just says that the result is "rounded". That is, at best, sloppy terminology. The C++ language definition says that integer division truncates. (Formally, it discards the fractional part). – Pete Becker Aug 07 '13 at 16:49

score 1 · Answer 5 · answered Aug 07 '13 at 16:34

1

(1/31536000) will yield 0 that multiplied by any number would be 0. Make atleast one of numerator or denominator float (like 1.0 or 31536000.0)

answered Aug 07 '13 at 16:34

Anshul

666
1
5
18

score 1 · Answer 6 · answered Aug 07 '13 at 16:35

1

883102.0 / 31536000 will do just what you want.

answered Aug 07 '13 at 16:35

Pete Becker

72,338
6
72
157

score 1 · Answer 7 · answered Aug 07 '13 at 16:36

1

Because of the Integer-Division

(1/31536000)

the fractional digits get truncate and the result is "zero". You have to add a dot:

(1.0/31536000.0) or (1./31536000.)

answered Aug 07 '13 at 16:36

Zagatho

444
1
6
19

Math Calculation - float numbers

7 Answers7