Data sampler: Allow selecting 100% of data, shuffle it

Orange/widgets/data/owdatasampler.py

@@ -60,7 +60,7 @@ class Warning(OWWidget.Warning):
 
     class Error(OWWidget.Error):
         too_many_folds = Msg("Number of folds exceeds data size")
-        sample_larger_than_data = Msg("Sample must be smaller than data")
+        sample_larger_than_data = Msg("Sample can't be larger than data")
         not_enough_to_stratify = Msg("Data is too small to stratify")
         no_data = Msg("Dataset is empty")
 
@@ -88,7 +88,7 @@ def set_sampling_type_i():
         self.sampleSizePercentageSlider = gui.hSlider(
             gui.indentedBox(sampling), self,
             "sampleSizePercentage",
-            minValue=0, maxValue=99, ticks=10, labelFormat="%d %%",
+            minValue=0, maxValue=100, ticks=10, labelFormat="%d %%",
             callback=set_sampling_type(self.FixedProportion),
             addSpace=12)
 
@@ -265,7 +265,7 @@ def updateindices(self):
         else:
             assert self.sampling_type == self.Bootstrap
 
-        if not repl and size is not None and (data_length <= size):
+        if not repl and size is not None and (size > data_length):
             self.Error.sample_larger_than_data()
         if not repl and data_length <= num_classes and self.stratify:
             self.Error.not_enough_to_stratify()
@@ -386,7 +386,12 @@ def __call__(self, table):
             o[sample] = 0
             others = np.nonzero(o)[0]
             return others, sample
-        if self.stratified and table.domain.has_discrete_class:
+        if self.n == len(table):
+            rgen = np.random.RandomState(self.random_state)
+            sample = np.arange(self.n)
+            rgen.shuffle(sample)
+            return np.array([], dtype=int), sample
+        elif self.stratified and table.domain.has_discrete_class:
             test_size = max(len(table.domain.class_var.values), self.n)
             splitter = skl.StratifiedShuffleSplit(
                 n_splits=1, test_size=test_size,

Orange/widgets/data/tests/test_owdatasampler.py

@@ -100,6 +100,22 @@ def test_bigger_output_warning(self):
         self.set_fixed_sample_size(3, with_replacement=True)
         self.assertTrue(self.widget.Warning.bigger_sample.is_shown())
 
+    def test_shuffling(self):
+        self.send_signal('Data', self.iris)
+
+        self.set_fixed_sample_size(150)
+        self.assertFalse(self.widget.Warning.bigger_sample.is_shown())
+        sample = self.get_output("Data Sample")
+        self.assertTrue((self.iris.ids != sample.ids).any())
+        self.assertEqual(set(self.iris.ids), set(sample.ids))
+
+        self.select_sampling_type(self.widget.FixedProportion)
+        self.widget.sampleSizePercentage = 100
+        self.widget.commit()
+        sample = self.get_output("Data Sample")
+        self.assertTrue((self.iris.ids != sample.ids).any())
+        self.assertEqual(set(self.iris.ids), set(sample.ids))
+
     def set_fixed_sample_size(self, sample_size, with_replacement=False):
         """Set fixed sample size and return the number of gui spin.
 

doc/visual-programming/source/widgets/data/datasampler.md

@@ -27,6 +27,8 @@ The **Data Sampler** widget implements several data sampling methods. It outputs
    input dataset.
 4. Press *Sample Data* to output the data sample.
 
+If all data instances are selected (by setting the proportion to 100 % or setting the fixed sample size to the entire data size), output instances are still shuffled.
+
 Examples
 --------