(Mis)use of the is_nan_or_na() mapper function

[Paul Laub]

Dear all, 

Am I missing something or is there a problem with the function 
is_nan_or_na? I am interested in using it to identify missing 
data elements (variously marked NaN or NA -- I know, NA is 
preferable). 

It appears that is_nan_or_na() does not detect NAs as this
transcript shows. I am using Octave 2.1.42 in Windows 2000. 

    184 >> which is_nan_or_na
    is_nan_or_na is a built-in mapper function 
    185 >> p = 1:10;
    186 >> p(3) = NaN;
    187 >> p(5) = NA;
    188 >> isna(p) | isnan(p)
    ans =

      0  0  1  0  1  0  0  0  0  0     # what I expect

    189 >> is_nan_or_na(p)
    ans =

      0  0  1  0  0  0  0  0  0  0     # not what I expect

    190 >> is_nan_or_na(NaN)
    ans = 1
    191 >> is_nan_or_na(NA)            # well ... 
    ans = 0

The workaround is obvious; it is command 188 above. But what gives
with is_nan_or_na()? 

Paul Laub 

P.S. A web search turned up this 2002 discussion issues. 

    http://www.octave.org/mailing-lists/help-octave/2002/581

What caught my eye is that in that exchange the function is 
named with different case, ie., is_NaN_or_NA. I know that
Octave is case-sensitive. 

P.P.S. Having NA is a good idea but how is it implemented?
How do you keep it separate from NaN and Inf and still use 
IEEE style floating point math? I notice that sprintf("%f", NA)
returns "NaN".



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------