How to take only 3 digits after float for calculation?

[Izzy]

Say for eg. you have a value stored in a float variable.

Say I stored

float number = 2.456;

But it will stored

float number = 2.4562134654789; random bunch of numbers after the assigned valaue.
Now say I just want to use 2.456 from this above value for calculation, how do you do that?

I know you can do by multiplying by 1000 then converting to int....and so on..

But isn't there any other easy way?

Thanks in advance

[Ttelmah]

Basically, no.
The float doesn't actually store 2.456. It stores a scaled binary representation of the value, and separately the scaling applied. Unfortunately, there is then no direct route from this value to a decimal 'number of digits'. You can as you say, multiply by 1000, and store as an integer. However remember that you should add 0.0004999 to the value (for +ve numbers) to get normal 4/5 rounding, or you will be always trimming 'downwards'.
Remember that you can use the %w printf formatter to output such scaled values directly, with the decimal point inserted, and that it is much faster if you can work with such values for the arithmetic itself.
For FP values there is normally not any reason to throw away accuracy. If you are going to do this, it should be the last thing done, or you increase the errors as arithmetic proceeds. For output, you always have the field width specifier, so the 'extra' characters need not be seen.

Best Wishes

[Izzy]

One more question:

When you print to a serial LCD, is there a way you can reserve space for ...say for eg 2.3 precision float? meaning 2 digit and 3 precision ...like 35.123

say if the print should start after 5th block on the 20 x 4 LCD.

_ _ _ _ _ 3 5 . 1 2 3

now I dont want the number to print

_ _ _ _ _ 2 . 2 2 3 if the number is 1 digit with 3 precision

I am trying to get something like

_ _ _ _ _ _ 2 . 2 2 3 ...Notice that the 6th block is empty, so that the decimal remains aligned.


Is it possible?


I tried using "%3.2f" but it only displays 2 digit with 3 precision , it does not leave the block empty if I am trying to print 1 float 3 precision float.


Does anyone knows about this?

[PCM programmer]

Do a test with an if() statement. If the number is less than 10.000
then print a space first. Then display the number.

[Izzy]

Thanks for the reply.

If there were only couple of values to be displayed, I would have done that.

But I have almost 30 - 40 values to be displayed, so I am not sure if it would be appropiate. I will see.

[Ttelmah]

%3.2, shouldn't give '3 precision'. The number after the decimal, is the number of digits wanted after the decimal point (should be '3' for you), while the number in front, is the _total_ field width (must also include the decimal point, and the digits after the decimal). Normally, the specifier you would want, would be % 6.3f. Note the 'space'. However (unfortunately), CCS does not support this... :-(
However the number of things you have to output, doesn't mean you can't use a test. Also, you only seem to be talking about possible numbers up to 99.999 max (since this is the space you are allowing).
Generate your own output routine, like

[Douglas Kennedy]

Ttelmah is right in mentioning that there is nothing untoward occurring with

[Morcilla]

i recomend to use: