ENH:read_html() handles tables with multiple header rows #13434

doc/source/io.rst

@@ -2232,9 +2232,10 @@ Read a URL and match a table that contains specific text
    match = 'Metcalf Bank'
    df_list = pd.read_html(url, match=match)
 
-Specify a header row (by default ``<th>`` elements are used to form the column
-index); if specified, the header row is taken from the data minus the parsed
-header elements (``<th>`` elements).
+Specify a header row (by default ``<th>`` or ``<td>`` elements located within a 
+``<thead>`` are used to form the column index, if multiple rows are contained within
+``<thead>`` then a multiindex is created); if specified, the header row is taken
+from the data minus the parsed header elements (``<th>`` elements).
 
 .. code-block:: python
 

doc/source/whatsnew/v0.20.0.txt

@@ -309,6 +309,7 @@ Other Enhancements
 - ``pandas.tools.hashing`` has gained a ``hash_tuples`` routine, and ``hash_pandas_object`` has gained the ability to hash a ``MultiIndex`` (:issue:`15224`)
 - ``Series/DataFrame.squeeze()`` have gained the ``axis`` parameter. (:issue:`15339`)
 - ``DataFrame.to_excel()`` has a new ``freeze_panes`` parameter to turn on Freeze Panes when exporting to Excel (:issue:`15160`)
+- ``pd.read_html()`` parses multiple header rows, creating a multiindex header. (:issue:`13434`).
 - HTML table output skips ``colspan`` or ``rowspan`` attribute if equal to 1. (:issue:`15403`)
 
 - ``pd.TimedeltaIndex`` now has a custom datetick formatter specifically designed for nanosecond level precision (:issue:`8711`)

pandas/io/html.py

@@ -355,9 +355,12 @@ def _parse_raw_thead(self, table):
         thead = self._parse_thead(table)
         res = []
         if thead:
-            res = lmap(self._text_getter, self._parse_th(thead[0]))
-        return np.atleast_1d(
-            np.array(res).squeeze()) if res and len(res) == 1 else res
+            trs = self._parse_tr(thead[0])
+            for tr in trs:
+                cols = lmap(self._text_getter, self._parse_td(tr))
+                if any([col != '' for col in cols]):
+                    res.append(cols)
+        return res
 
     def _parse_raw_tfoot(self, table):
         tfoot = self._parse_tfoot(table)
@@ -591,9 +594,17 @@ def _parse_tfoot(self, table):
         return table.xpath('.//tfoot')
 
     def _parse_raw_thead(self, table):
-        expr = './/thead//th'
-        return [_remove_whitespace(x.text_content()) for x in
-                table.xpath(expr)]
+        expr = './/thead'
+        thead = table.xpath(expr)
+        res = []
+        if thead:
+            trs = self._parse_tr(thead[0])
+            for tr in trs:
+                cols = [_remove_whitespace(x.text_content()) for x in
+                        self._parse_td(tr)]
+                if any([col != '' for col in cols]):
+                    res.append(cols)
+        return res
 
     def _parse_raw_tfoot(self, table):
         expr = './/tfoot//th|//tfoot//td'
@@ -615,19 +626,17 @@ def _data_to_frame(**kwargs):
     head, body, foot = kwargs.pop('data')
     header = kwargs.pop('header')
     kwargs['skiprows'] = _get_skiprows(kwargs['skiprows'])
-
     if head:
-        body = [head] + body
-
+        rows = lrange(len(head))
+        body = head + body
         if header is None:  # special case when a table has <th> elements
-            header = 0
+            header = 0 if rows == [0] else rows
 
     if foot:
         body += [foot]
 
     # fill out elements of body that are "ragged"
     _expand_elements(body)
-
     tp = TextParser(body, header=header, **kwargs)
     df = tp.read()
     return df

pandas/tests/io/test_html.py

@@ -760,6 +760,18 @@ def test_keep_default_na(self):
         html_df = read_html(html_data, keep_default_na=True)[0]
         tm.assert_frame_equal(expected_df, html_df)
 
+    def test_multiple_header_rows(self):
+        # Issue #13434
+        expected_df = DataFrame(data=[("Hillary", 68, "D"),
+                                      ("Bernie", 74, "D"),
+                                      ("Donald", 69, "R")])
+        expected_df.columns = [["Unnamed: 0_level_0", "Age", "Party"],
+                               ["Name", "Unnamed: 1_level_1",
+                                "Unnamed: 2_level_1"]]
+        html = expected_df.to_html(index=False)
+        html_df = read_html(html, )[0]
+        tm.assert_frame_equal(expected_df, html_df)
+
 
 def _lang_enc(filename):
     return os.path.splitext(os.path.basename(filename))[0].split('_')