Rounding REAL numbers

[wowski]

I can’t find an Fortran function that rounds a real number to a given decimal place (such as ROUND(x,y) in VB). The only way I’ve found of doing it which works but is ugly is, for the case of wanting to round to the third decimal:



dummyVar = 123.456789

dummyVar = TRUNC(dummyVar * 1000.0) / 1000.0



There must be a better way? Any ideas?

 

[johnhors]

wowski

There is no elegant method to do this. Your TRUNC function is not part of the Fortran standard and I have never seen this in any of the Fortran compilers that I have used. Instead you can use INT to remove all the decimal places and FLOAT to return the variable to a real. It is standard practice to round numbers up by adding 0.5 before taking the integer.

My code to do this is:-

PROGRAM TRUNCATE

REAL*8 A,B

A=12345.6789

B=FLOAT (INT(A * 1000.0 + 0.5)) / 1000.0

WRITE(*,*)'A,B =',A,B

STOP
END

However why would you need to do this? It is customary to use an appropriate format to write numbers out with the required number of decimal places.

 

[SErgini0]

I usually put restriction on the output.
example:

11111 format(1f8.1)
write(*,11111)dummyVar