Round function and decimal number

[manuelb]

From the old TOL forums:

Hi.
I think the Round function should have a second argument (optional) that indicates the
number of decimals you want it to round the number to, defaulting to zero.

Now the function is defined:

Real Round(Real r)

--> Real Round(4.8369) = 5;

I propose:

Real Round(Real x [, Real p=0]):
--> Real Round(4.8369) = 5;
--> Real Round(4.8369, 2) = 4.84;

Regards.

[rcsoto]

Since C++ does not implement the Round() with two parameters, the mechanism
can be used that is used at the moment in tol.
If "p" is distinct from zero:

Round(n*10^{p) / (10}p)

Example:

Real n = 1234.56789;
Real p1 = 2;
Real p2 = -2;

Real r1 = (Round(n*10^{p1)) / (10}p1);
Real r2 = (Round(n*10^{p2)) / (10}p2);

I hope that it serves as help.

[danirus]

Implementd the enhancement proposed by Ivan consistent in provide a second argument to
Round function of Real type, telling that the result must have a number of decimal digits
instead of return an integer.
This new argument is optional.

[Víctor de Buen Remiro]

In next code

Real n = 14.8369;
Real p1 = 1;
Real p2 = -1;
Real r1 = (Round(n*10^{p1)) / (10}p1);
Real r2 = (Round(n*10^{p2)) / (10}p2);
Real r1_= Round(n,p1);
Real r2_= Round(n,p2);

r1 is equal to r1 and r2_ should be equal to r2 but it is not happens thus.

[Víctor de Buen Remiro]

The correct value is r2 = 10, not r2_=14.8369
It's to say the number of decimals should be extended with negative values in the same way that positive ones to round a value to decenes, hundreds, etc.

[Víctor de Buen Remiro]

Problem has been solved in CVS