Where does the garbage in a floating point variable come from in Delphi?

Question

Delphi

There is a text field where I type something like 340.788
Then I take StrToFloat from the text into the field into the single variable, then again FloatToStr - and I get - 340.787994384766.
If you round it up, and then add, say, 100, all the same, garbage is added in the lower digits.

Answer 1

Because the number 340.788 cannot be represented exactly in floating point. None: neither single, nor double, nor extended. And FloatToStr on standard settings assumes the accuracy of either double or extended.

Answer 2

It is taken because of the peculiarities of storing fractional numbers.

Answer 3

Try Double

Answer 4

Or use Round()

Answer 5

Probably all the same, fractions are stored in binary and not decimal still at school, this all goes through

Answer 6

I solved the problem a little strange - I removed the integer part, multiplied the fractional one to the accuracy I needed (* 1000) after the sign, rounded it up to the integer, removing the fractional part .. I just had to get a number equal to thousandths of the value.