Variable floating-point precision format string

I'm trying to print floating point numbers as percentages, and I'd like for the number of digits after the decimal place to vary as needed. Currently I have:

fmt.Printf("%.2f%%\n", 100*(value/total))

The problem is that if my percentage is, say, exactly 50, I will get the following output:

50.00%

While what I want is to get:

50%

Is there any way for the format string to indicate that a maximum of 2 digits of precision should be used, but only if needed?

There's no direct solution with the fmt package.

But you can remove the dot and zeros at end with a regular expression:

r, _ := regexp.Compile(`\.?0*$`)
fmt.Printf("%s%%\n", r.ReplaceAllString(fmt.Sprintf("%.2f", 100*(value/total)),""))

Bonus: the same regex works for any number of trailing zeros.

Side note: You'll display 50.0041 the same way than 50, which might be a little misleading.

There's no way to do that inside fmt with e.g. another flag or what have you. You'll have to write out the logic yourself. You could do something like:

var final string
doubledecimal := fmt.Sprintf("%.2f", 100*value/total)
if doubledecimal[len(doubledecimal)-2:] == "00" {
	final = doubledecimal[:len(doubledecimal)-3]
} else {
	final = doubledecimal
}
fmt.Printf("%s%%\n, final)

You could similarly use strings.Split to split on the decimal point and work from there.

You could even adjust this to turn 50.10% into 50.1%.

doubledecimal := fmt.Sprintf("%.2f", 100*value/total)
// Strip trailing zeroes
for doubledecimal[len(doubledecimal)-1] == 0 {
    doubledecimal = doubledecimal[:len(doubledecimal)-1]
}
// Strip the decimal point if it's trailing.
if doubledecimal[len(doubledecimal)-1] == "." {
    doubledecimal = doubledecimal[:len(doubledecimal)-1]
}
fmt.Printf("%s%%\n", doubledecimal)

One way could be to have an if statement controlling the print output, i.e. if the result is cleanly divisible by 1 (result%1 == 0) then print the result to no decimal places. Otherwise print to .2f as you've done above. Not sure if there is a shorter way of doing this, but I think this should work.