[Octave-bug-tracker] [bug #54275] repelem error if a repeat instruction is a vector ends with 0

[anonymous]

Follow-up Comment #1, bug #54275 (project octave):

The first solution in the original submission is a little ugly.
If we didn't want to use the second solution (using "repelems"),
here is a third solution.

--- a/scripts/general/repelem.m
+++ b/scripts/general/repelem.m
@@ -299,20 +299,19 @@
 
   else
     ## This works for a row or column vector.
-    ## idx2 output will be a row vector.
 
     ## Get ending position for each element item.
     idx_temp = cumsum (v);
 
-    ## Row vector with enough space for output
-    idx(1:idx_temp(end)) = 0;
-
     ## Set starting position of each element to 1.
-    idx(idx_temp(1:end-1) + 1) = 1;
+    idx(idx_temp + 1) = 1;
 
     ## Set starting position of each element to 1.
     idx(1) = 1;
 
+    ## Row vector with proper length for output
+    idx = idx(1:idx_temp(end));
+
     ## with prepared index
     idx = (find (v != 0))(cumsum (idx));
 


Besides, in fact, in the subfuntion prepareIdx, the variable "idx" is actually
a row vector 
except for the last line "idx = (find (v != 0))(cumsum (idx))".
The problem is that "find(v!=0)" might be a row or column vector according to
the shape of "v".
If we really wanted the output "idx" to be a row vector, 
we could change that line into "idx = find (v != 0)(:).'(cumsum (idx))".

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?54275>

_______________________________________________
  Message sent via Savannah
  https://savannah.gnu.org/