Home »
Public Forums »
archive »
Re: Calculating mean (was mean and sdev)
| Re: Calculating mean (was mean and sdev) [message #11037] |
Thu, 05 March 1998 00:00 |
nospam
Messages: 21
Registered: November 1997
|
Junior Member |
|
|
In article <34FE5D52.70D510F8@linmpi.mpg.de> Kevin Ivory <Kevin.Ivory@linmpi.mpg.de> writes:
> Kevin Spencer wrote:
>> they get rid of stdev in the first place? And calculating the mean
>> is trivial; just use total(x)/n_elements(x).
>
> It is not as trivial as that. If your array contains NaNs, you will
> always get a NaN as a result. In many cases you will want to have
> the mean of the finite values.
I just looked at the source for moment in IDL 5.0.2, and it appears
that it does not correctly handle NaNs.
TypeX = SIZE(X)
nX = TypeX[TypeX[0]+2]
Mean = DATOTAL(X, Double = Double) / nX
the function DATOTAL just calls total() and massages the return type.
The documentation for total() says that you have to set /NAN to have
it treat NaNs as missing data, and the call to total does not do that,
nor does it remove the NaNs from the count of elements.
scott
--
Scott Stuart
stuart at ll mit edu
|
|
|
|
Current Time: Sat Oct 11 09:27:30 PDT 2025
Total time taken to generate the page: 1.36198 seconds
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
