objects, subrefs and numel

[CdeMills]

Hello,

Octave 3.8 doc states that:
Built-in Function: numel (A, IDX1, IDX2, ...)
     This method is also called when an object appears as lvalue with
     cs-list indexing, i.e., 'object{...}' or 'object(...).field'.

I recently implement the multi-argument form of numel for dataframes
objects:
switch nargin
    case 1
      %# use the internal structure fields to get the total number of
element
      resu = prod (df._cnt);
    otherwise
      %# do the subsreferencing and call numel on the resulting
sub-dataframe
      S.type = '()';
      S.subs = varargin;
      resu = numel (subsref (df, S));

This means that constructions as df{} results in TWO calls to subrefs:
- one in the context of numel(), generating an object from df(...)
- one proper, evaluating df{...}
Both calls to subrefs results in objects of the same size. But the
construction is quite complex. Would it be possible to change the logic as
1) do the sub-referencing
2) call the single-argument form of numel() on the returned object ?

This would avoid this double operation, as the object returned from the
first call is just inspected and throw away.

Regards

Pascal







--
View this message in context: 
http://octave.1599824.n4.nabble.com/objects-subrefs-and-numel-tp4668682.html
Sent from the Octave - Maintainers mailing list archive at Nabble.com.