bradipo

as they're easy to find, I read several threads about the round-off problem everybody who ever trafficked around with int and float objects is well aware of. So I know Max isn't bad at maths and all.

That said, my specific problem consists in a Max project I'm writing that uses the Max window to give a lot of information with float numbers I'd like to be printed with a given precision. If I used only int numbers, the given information would not be as accurate as I need it to be, and using float numbers means the information gets unreadable because of seeing a series of 36.270000002 instead of 36.27 (for instance). So, did anybody figure a method to stop Max printing all those unneeded decimals, or is the only way the decomposition of my numbers in two parts, tricking around with the sprintf object ?

any-chance-to-get-a-number-like-3-814-printed-in-the-max-window

sprintf should format a number the way you want.

36.270000002 is 36.270000002 and if you want it to read 36.27 you want to

print something different from what you have.

your idea to simply eliminate the last 2-8 digits seems right, [sprintf] will work

36.270000002 is 36.270000002 and if you want it to read 36.27 you want to
print something different from what you have.

your idea to simply eliminate the last 2-8 digits seems right, [sprintf] will work
but [expr] will work too.

Thanks for the idea. Something strange happens when I use the expr object. I tried the very simple formula expr int ($f1 * 1000.) / 1000. which should work, but it does only sometimes. In case this expression is given a number as 4.72953168, the number will be truncated down to 4.729. But if it's given 4.7290003 or 4.7299997, the numbers rest exactly as they were.

It seems that Max "knows" that in saying 4.7290003 it really means to say 4.729, and then [expr] doesn't work. That's a pity. I'll have to slice my numbers in two parts and use [sprintf].

Thanks for the idea. Something strange happens when I use the expr object. I tried the very simple formula expr int ($f1 * 1000.) / 1000. which should work, but it does only sometimes. In case this expression is given a number as 4.72953168,  the number will be truncated down to 4.729. But if it's given 4.7290003 or 4.7299997, the numbers rest exactly as they were.

i'm not sure why you're even bothering trying to use expr to format the string representation of a floating point number. it simply doesn't work that way.

also, I don't know why you need to "slice your numbers into two parts." If you want 3 decimal digits, just use the format string 

having said that, sprintf has a problem when formatting one number. after the formatted string leaves the sprintf object, max has a habit of magically turning it back into a floating point number, so if you want to push it through a print, you're going to get a bunch of trailing zeros as print does what it does when it prints numbers.

the solution (the best one I can come up with, anyhow) is to have a more complicated format string that max won't mess with, like 

also, I don't know why you need to "slice your numbers into two parts." If you want 3 decimal digits, just use the format string %.3f.

the solution (the best one I can come up with, anyhow) is to have a more complicated format string that max won't mess with, like "the answer is %.3f"


This will prevent sprintf from turning it back into a float.

easy!
Use the symout argument.
[sprintf symout %.3f]

This will prevent sprintf from turning it back into a float.


doh. i forgot about the symout argument.


Thanks again for your help ! Mudang's idea worked. I believe that you've to know C language in order to know about this %.3f format, and I don't. That's why I didn't manage before.

Any chance to get a number like 3.814 printed in the Max window ?