Merge 78cb444 into 6aa94b2

q2_feature_table/_merge.py

@@ -54,7 +54,7 @@ def merge(tables: biom.Table,
                          (overlap_method, ', '.join(overlap_methods())))
 
 
-def _merge_feature_data(data: pd.Series) -> pd.Series:
+def _merge_feature_data(data):
     data = iter(data)
     result = next(data)  # There is always at least 1
     for d in data:
@@ -66,5 +66,11 @@ def merge_seqs(data: pd.Series) -> pd.Series:
     return _merge_feature_data(data)
 
 
-def merge_taxa(data: pd.Series) -> pd.Series:
-    return _merge_feature_data(data)
+def merge_taxa(data: pd.DataFrame) -> pd.DataFrame:
+    data = _merge_feature_data(data)
+    # merge orders columns alphabetically; Taxon must be first header column
+    # as defined here: https://github.com/qiime2/q2-types/blob/
+    # 067d83e2aefe98674433e95162336fb5b9d96474/q2_types/feature_data/
+    # _format.py#L97
+    data = data[data.columns.drop('Taxon').insert(0, 'Taxon')]
+    return data

q2_feature_table/tests/test_merge.py

@@ -286,15 +286,20 @@ def test_merge_seqs(self):
 class MergeFeatureTaxonomyTests(unittest.TestCase):
     # More extensive testing is performed in MergeFeatureDataTests, which
     # tests the shared private API.
+    # This tests a specifically FeatureData[Taxonomy]-like dataframe
+    # and ensures delivery in valid format (Taxon column first)
 
     def test_merge_taxa(self):
         # this test calls the public API directly
-        d1 = pd.Series(['a;b;c;d', 'a;b;c;e'], index=['f1', 'f2'])
-        d2 = pd.Series(['a;b;c;d', 'a;b;c;e'], index=['f1', 'f3'])
+        d1 = pd.DataFrame([('a;b;c;d', '1.0'), ('a;b;c;f', '0.7')],
+                          index=['f1', 'f2'], columns=['Taxon', 'Confidence'])
+        d2 = pd.DataFrame([('1.0', 'a;b;c;g'), ('1.0', 'a;b;c;e')],
+                          index=['f1', 'f3'], columns=['Confidence', 'Taxon'])
         obs = merge_taxa([d1, d2])
-        exp = pd.Series(['a;b;c;d', 'a;b;c;e', 'a;b;c;e'],
-                        index=['f1', 'f2', 'f3'])
-        pdt.assert_series_equal(obs, exp)
+        exp = pd.DataFrame(
+            [('a;b;c;d', '1.0'), ('a;b;c;f', '0.7'), ('a;b;c;e', '1.0')],
+            index=['f1', 'f2', 'f3'], columns=['Taxon', 'Confidence'])
+        pdt.assert_frame_equal(obs, exp)
 
 
 if __name__ == "__main__":