Fix isin bug (#16049)

h2o-py/h2o/frame.py

@@ -1845,7 +1845,7 @@ def isin(self, item):
         """
         if is_type(item, list, tuple, set):
             if self.ncols == 1 and (self.type(0) == 'str' or self.type(0) == 'enum'):
-                return self.match(item, nomatch=0)
+                return self.match(item, nomatch=0) > 0
             else:
                 return functools.reduce(H2OFrame.__or__, (self == i for i in item))
         else:

h2o-py/tests/testdir_misc/pyunit_isin.py

@@ -13,9 +13,22 @@ def isin_check():
   assert not (cars.isin(["AMC Gremlin","AMC Concord DL"]) == cars.isin("AMC Gremlin")).all()
   assert (cars.isin(["AMC Gremlin","AMC Concord DL",6]) == cars.isin("AMC Gremlin") + cars.isin("AMC Concord DL") 
           + cars.isin(6)).all()
+
+
+def test_isin_issue_16043():
+    frame = h2o.upload_file(pyunit_utils.locate("smalldata/titanic/titanic_expanded.csv"))
+    column = "cabin"
+    levels = frame[column].levels()[0][:-10]
+    assert len(levels) > 10
+    print(frame[column].isin(levels).unique())
+    # expecting a binary vector as in pandas
+    assert frame[column].isin(levels).unique().nrows == 2
+    assert frame.as_data_frame(use_pandas=True)[column].isin(levels).nunique() == 2
 
 
 if __name__ == "__main__":
   pyunit_utils.standalone_test(isin_check)
+  pyunit_utils.standalone_test(test_isin_issue_16043)
 else:
   isin_check()
+  test_isin_issue_16043()