CINT rounding

[Rodney Hicks]

From the PBWin 9.0 help on the CINT-CWRD group:

In the case of the functions that convert to integer-class values, the fractional part of the number is rounded to the nearest even number. For example, CINT(1.5) returns 2, CINT(.5) returns 2, and CLNG(-0.6) returns -1.

It seems that the last two examples are not in sync with the statement. It seems pretty obvious to me but I thought I'd post just in case I've had my head buried in numbers too long to understand words before sending it in to support.
I would think that if the numbers are being rounded to the nearest even integer that the last two examples would return 0, for those of us that consider 0 an even integer.

After testing, CINT(0.5) returns 0, CINT(-0.6) returns -1, CINT(0.6) returns 1.

[John Gleason]

You're not hallucinating or anything. It definitely could use a support rewrite. The tricky part is with the ".5" instance: eg. 1.5 rounds up to 2 because odd (referring to the 1st number left of the decimal pt.) rounds up, and 2.5 rounds down to 2 because even rounds down in what's called unbiased or bankers rounding.

[Rodney Hicks]

This statement from the PBWin 8 manual is as follows:
"In the case of the functions that convert to integer-class values, the result is rounded up if it has a fractional component of 0.5 or greater. For example, CINT(0.7) returns 1 and CLNG(-0.6) returns -1."

This is different than the bankers rounding in PBWin 9.00.

[Michael Mattias]

See also: CEIL() and FIX() functions. (especially if dealing with signed numbers)

[John Gleason]

Good catch on those PB8 docs too Rod. The function works the same in 8 and 9--both use banker's rounding--but the docs need correction in both. My docs in ver. 8 say For example, CINT(1.5) returns 1 which isn't correct.

[Rodney Hicks]

And in PBCC 4 the help file states:

In the case of the functions that convert to integer-class values, the result is rounded down if it has a fractional component of 0.5 or less. For example, CINT(1.5) returns 1, and CLNG(-0.6) returns -1.

And the manual says "rounded up if it has a fractional component of 0.5 or greater.

I don't have PBCC 5 yet so I can't check that, so I'll leave that to someone that does have it.

Not a much used function so it probably escaped notice.