http://octave.sourceforge.net/doc/f/values.html
>From values() documentation:
"Return the different values in a column vector, arranged in ascending
order."
If source vector for values() contains NaN values, they appear last in the
returned vector.
This clearly appear in the source of the function which contain the following
code:
if (i) <-- "i" flags that NaNs where found
v = [v; NaN]; <-- add to the vector the NaN at the end.
table() function relies on values() function
http://octave.sourceforge.net/doc/f/table.html
so if there are NaN values they are aggregated in the last rows/columns.
Is this behavior a part of the spec? Can one rely that NaNs will always appear
at the end, even in future releases?
Is it a convention to treat NaN as the highest value? These are softwares (eg,
SAS) that treat missing values as the lowest (smaller than the maximum negative
value).
Please highlight this issue more exlicitly in the documentation.
_________________________________________________________________
Insert movie times and more without leaving Hotmail®.
http://windowslive.com/Tutorial/Hotmail/QuickAdd?ocid=TXT_TAGLM_WL_HM_Tutorial_QuickAdd1_052009
------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensing option that enables
unlimited royalty-free distribution of the report engine
for externally facing server and web deployment.
http://p.sf.net/sfu/businessobjects
_______________________________________________
Octave-dev mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/octave-dev
