You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
An ncl-talk user on 12/9/2019 reported a problem using in which they would get this error message:
fatal:simpne: Must have three or more non-missing values.
When in fact their data had no missing values in the x variable. Indeed, this same message results when running the 1st two 1D examples in the documentation page for simpne().
That error message originates in .../ncl/ni/src/lib/nfp/simpsonW.c. The git log shows that back in 2016, an explicit test for more than 3 non-missing vals was added. At around lines 427-428 (at time of writing), there's this:
nvalid = size_leftmost - nmiss;
if(nvalid < 3) {
Following the code logic, if the x input array is 1D, "size_leftmost" will always be 1, and thus "nvalid" always < 3. I'm not certain about an ND input array and I don't really understand the computation of size_leftmost for the ND case. However, I think the calculation for "nvalid" should be something like:
nvalid = npts - nmiss
The text was updated successfully, but these errors were encountered:
An ncl-talk user on 12/9/2019 reported a problem using in which they would get this error message:
fatal:simpne: Must have three or more non-missing values.
When in fact their data had no missing values in the x variable. Indeed, this same message results when running the 1st two 1D examples in the documentation page for simpne().
That error message originates in .../ncl/ni/src/lib/nfp/simpsonW.c. The git log shows that back in 2016, an explicit test for more than 3 non-missing vals was added. At around lines 427-428 (at time of writing), there's this:
Following the code logic, if the x input array is 1D, "size_leftmost" will always be 1, and thus "nvalid" always < 3. I'm not certain about an ND input array and I don't really understand the computation of size_leftmost for the ND case. However, I think the calculation for "nvalid" should be something like:
nvalid = npts - nmiss
The text was updated successfully, but these errors were encountered: