CODEWITHSUNDEEP
c++floating-point
RuQu
RuQu

Reputation: 253

Why do my floats insist on staying integers?

I am reading in several integers, which represent the year, julian day, hour, and minutes. I am trying to convert them to fractional days.

int YYYY, JJJ, HH, MM;

float datenumber = (YYYY*360.0f)+(JJJ*1.0f)+((HH*1.0f)+(MM/60.0f))/24.0f;

Using the values of 2001, 083, 22, 32 I should get a result of 724043.939. Instead I get 724044.

I have all the ints cast as floats. Why do they stay as integers?

edit Yes, I was displaying the output with cout. setprecision resolved the problem, thank you.

Upvotes: 3

Views: 219

Answers (5)

Dan F
Dan F

Reputation: 17732

Judging from your input and output values in the example you provided in Benjamin's answer, I am willing to bet that the issue is just floating point precision. Just because you declare something as a float, doesn't mean it can handle any size number. The problem lies in the fact that your year component (1200*360=432000) is WAY larger than your MM component(5/60/24=.00347222). The decimal portion from the MM part is just ignored because floats aren't that accurate. Try adding .000001 to 123456789 as a float and see what your output is

Upvotes: 0

Jacob
Jacob

Reputation: 34601

You need to do TWO things:

  1. Set your precision to the max (cout.setprecision(16))
  2. Convert to double (double dateNumber and YYYY*360.0)

Upvotes: 1

Benjamin Lindley
Benjamin Lindley

Reputation: 103693

The problem is not in your number or your calculation. It is only in the way you are displaying it. cout has decided that 6 digits is enough for you. Use setprecision if you want more.

std::cout << std::setprecision(10) << datenumber;

demo

Upvotes: 3

user1065734
user1065734

Reputation:

The following may be the reason- 1.May be you are printing them wrongly. 2.Dividing by an integer,use float instead.

Example-Instead of dividing by 200 divide by 200.0

Upvotes: 0

wvm2008
wvm2008

Reputation: 3034

Unless you've got some more code that does the casting elsewhere, you're doing your cast after you've done the equation, so you are getting a float back, but using the ints in the equation before your cast drops the decimals.

Why not declare them as floats initially?

float YYYY, JJJ, HH, MM;

Upvotes: 0

Related Questions