# HG changeset patch
# User John W. Eaton <jwe@octave.org>
# Date 1267378336 18000
# Node ID 1f11fabfa34957118b2a7c3a118f2e9e1c8cbd51
# Parent  634e182d34e461e091c60dab5e54a77d43582370
improve performance of unique for sparse arrays

diff --git a/scripts/ChangeLog b/scripts/ChangeLog
--- a/scripts/ChangeLog
+++ b/scripts/ChangeLog
@@ -1,3 +1,10 @@
+2010-02-28  John W. Eaton  <jwe@octave.org>
+
+	* set/unique.m: If the argument is sparse and we are not
+	operating on rows and we don't need indices, convert nonzero
+	elements to a full matrix and work on that instead, converting
+	back to sparse when done.
+
 2010-02-28  John W. Eaton  <jwe@octave.org>
 
 	* set/unique.m: Return 0x1 arrays for empty arrays with some
diff --git a/scripts/set/unique.m b/scripts/set/unique.m
--- a/scripts/set/unique.m
+++ b/scripts/set/unique.m
@@ -74,6 +74,26 @@
     optrows = 0;
   endif
 
+  ## FIXME -- the operations
+  ##
+  ##   match = (y(1:n-1) == y(2:n));
+  ##   y(idx) = [];
+  ##
+  ## are very slow on sparse matrices.  Until they are fixed to be as
+  ## fast as for full matrices, operate on the nonzero elements of the
+  ## sparse array as long as we are not operating on rows.
+
+  ## FIXME -- unique is called when PKG_ADD files are parsed, but
+  ## issparse is not yet available because it is coming from a .oct
+  ## file?!?
+
+  if (exist ("issparse"))
+    if (issparse (x) && ! optrows && nargout <= 1)
+      y = unique ([0; (full (nonzeros (x)))], varargin{:});
+      return;
+    endif
+  endif
+
   if (optrows)
     n = size (x, 1);
     dim = 1;
@@ -132,7 +152,6 @@
     i(idx) = [];
   endif
 
-
 endfunction
 
 %!assert(unique([1 1 2; 1 2 1; 1 1 2]),[1;2])