BUG: Retain tz-aware dtypes with melt (#15785) (#20292)

Author: mroeschke

doc/source/whatsnew/v0.23.0.txt

@@ -896,6 +896,7 @@ Timezones
 - Bug in :func:`Timestamp.tz_localize` where localizing a timestamp near the minimum or maximum valid values could overflow and return a timestamp with an incorrect nanosecond value (:issue:`12677`)
 - Bug when iterating over :class:`DatetimeIndex` that was localized with fixed timezone offset that rounded nanosecond precision to microseconds (:issue:`19603`)
 - Bug in :func:`DataFrame.diff` that raised an ``IndexError`` with tz-aware values (:issue:`18578`)
+- Bug in :func:`melt` that converted tz-aware dtypes to tz-naive (:issue:`15785`)
 
 Offsets
 ^^^^^^^

pandas/core/reshape/melt.py

@@ -13,7 +13,9 @@
 
 import re
 from pandas.core.dtypes.missing import notna
+from pandas.core.dtypes.common import is_extension_type
 from pandas.core.tools.numeric import to_numeric
+from pandas.core.reshape.concat import concat
 
 
 @Appender(_shared_docs['melt'] %
@@ -70,7 +72,12 @@ def melt(frame, id_vars=None, value_vars=None, var_name=None,
 
     mdata = {}
     for col in id_vars:
-        mdata[col] = np.tile(frame.pop(col).values, K)
+        id_data = frame.pop(col)
+        if is_extension_type(id_data):
+            id_data = concat([id_data] * K, ignore_index=True)
+        else:
+            id_data = np.tile(id_data.values, K)
+        mdata[col] = id_data
 
     mcolumns = id_vars + var_name + [value_name]
 

pandas/tests/reshape/test_melt.py

@@ -212,6 +212,27 @@ def test_multiindex(self):
         res = self.df1.melt()
         assert res.columns.tolist() == ['CAP', 'low', 'value']
 
+    @pytest.mark.parametrize("col", [
+        pd.Series(pd.date_range('2010', periods=5, tz='US/Pacific')),
+        pd.Series(["a", "b", "c", "a", "d"], dtype="category"),
+        pd.Series([0, 1, 0, 0, 0])])
+    def test_pandas_dtypes(self, col):
+        # GH 15785
+        df = DataFrame({'klass': range(5),
+                        'col': col,
+                        'attr1': [1, 0, 0, 0, 0],
+                        'attr2': col})
+        expected_value = pd.concat([pd.Series([1, 0, 0, 0, 0]), col],
+                                   ignore_index=True)
+        result = melt(df, id_vars=['klass', 'col'], var_name='attribute',
+                      value_name='value')
+        expected = DataFrame({0: list(range(5)) * 2,
+                              1: pd.concat([col] * 2, ignore_index=True),
+                              2: ['attr1'] * 5 + ['attr2'] * 5,
+                              3: expected_value})
+        expected.columns = ['klass', 'col', 'attribute', 'value']
+        tm.assert_frame_equal(result, expected)
+
 
 class TestLreshape(object):