C++ multiply float by fraction

Question

just started C++ from a MATLAB background and getting confused.

float farenheit, celcius;
cin >> farenheit;
celcius = (farenheit - 32) * (5 / 9);
cout << "Temperature (c): " << celcius;

why does multiplying by 5/9 not work as expected, but this does?:

float farenheit, celcius;
cin >> farenheit;
celcius = ((farenheit - 32) * 5) / 9);
cout << "Temperature (c): " << celcius;

Thanks!

Look at this one https://stackoverflow.com/q/9455271/1741542 — Olaf Dietsche, Jun 20 '20 at 21:30
What does `5/9` produce on its own? Hint: They are both integers. — Ted Lyngmo, Jun 20 '20 at 21:30
Or even better this one https://stackoverflow.com/q/8906722/1741542 — Olaf Dietsche, Jun 20 '20 at 21:30
because `int/int` in C++ returns a `int` value, to solve this you can just append a `.0` to one of the 2 numbers `(5 / 9.0);` — Alberto Sinigaglia, Jun 20 '20 at 21:32
Some big differences with matlab: integer arithmetic, 0-based arrays. — stark, Jun 20 '20 at 21:32
Do you get a different result if you use the correct types, like `celcius = (farenheit - 32.f) * (5.f / 9.f);` ? — Ted Lyngmo, Jun 20 '20 at 21:33
@＿Static＿assert `(5.f / 9.f);` as I suggested would be quite sufficienct though. There's no need to bring in `double`s here. — Ted Lyngmo, Jun 20 '20 at 21:44
@TedLyngmo I'd argue that OP should switch to using `double` exclusively :) — Aykhan Hagverdili, Jun 20 '20 at 21:51
@＿Static＿assert Yes, that's often (nowadays) the wise thing to do. — Ted Lyngmo, Jun 20 '20 at 22:07

score 4 · Accepted Answer · edited Jun 20 '20 at 21:51

4

Thanks everyone,

C++ interprets 5 and 9 as int values so 5/9 is also an int. 5/9 = 0.566 which is truncated down to 0.

To fix this, append .0 or .f to the values be interpreted as a double or float respectively.

edited Jun 20 '20 at 21:51

Aykhan Hagverdili

answered Jun 20 '20 at 21:45

Was writing an answer myself, but I'm glad you figured it out before I could get it posted. Good work. – Jun 20 '20 at 21:47
1

Note: Appending `.0` makes the value into a `double` while `.f` makes it a `float`. You work with `float`s so, I'd stay with those to not waste CPU. – Ted Lyngmo Jun 20 '20 at 21:48
1

Thanks, @TedLyngmo. Amended my answer. – Helpmeplzimnotfrombath Jun 20 '20 at 21:51

score 1 · Answer 2 · answered Jun 20 '20 at 22:57

1

C++ considers 5 and 9 to be integers and the division is integer division, which means 5/9 = 0 (it returns the quotient).

So use 5.0 and 9.0 if you want them to be floating points.

answered Jun 20 '20 at 22:57

Harshit Agrawal

2 Answers2