ENH: add support for reading binary Excel files (#15914)

doc/source/whatsnew/v1.0.0.rst

@@ -212,6 +212,7 @@ Other enhancements
 - :meth:`DataFrame.to_markdown` and :meth:`Series.to_markdown` added (:issue:`11052`)
 
 - :meth:`DataFrame.drop_duplicates` has gained ``ignore_index`` keyword to reset index (:issue:`30114`)
+- :func:`read_excel` now accepts binary objects (:issue:`15914`)
 
 Build Changes
 ^^^^^^^^^^^^^

pandas/io/excel/_base.py

@@ -352,6 +352,8 @@ def __init__(self, filepath_or_buffer):
             self.book = self.load_workbook(filepath_or_buffer)
         elif isinstance(filepath_or_buffer, str):
             self.book = self.load_workbook(filepath_or_buffer)
+        elif isinstance(filepath_or_buffer, bytes):
+            self.book = self.load_workbook(BytesIO(filepath_or_buffer))
         else:
             raise ValueError(
                 "Must explicitly set engine if not passing in buffer or path for io."

pandas/tests/io/excel/test_readers.py

@@ -988,3 +988,14 @@ def test_conflicting_excel_engines(self, read_ext):
         with pd.ExcelFile("test1" + read_ext) as xl:
             with pytest.raises(ValueError, match=msg):
                 pd.read_excel(xl, engine="foo")
+
+    def test_excel_read_binary(self, engine, read_ext):
+        # GH 15914
+        expected = pd.read_excel("test1" + read_ext, engine=engine)
+
+        with open("test1" + read_ext, "rb") as f:
+            data = f.read()
+            f.close()
+
+        actual = pd.read_excel(data, engine=engine)
+        tm.assert_frame_equal(expected, actual)