Append Mode for ExcelWriter with openpyxl (pandas-dev#21251)

doc/source/whatsnew/v0.24.0.txt

@@ -8,6 +8,8 @@ v0.24.0
 New features
 ~~~~~~~~~~~~
 
+- ``ExcelWriter`` now accepts ``mode`` as a keyword argument, enabling append to existing workbooks when using the ``openpyxl`` engine (:issue:`3441`)
+
 .. _whatsnew_0240.enhancements.other:
 
 Other Enhancements

pandas/io/excel.py

@@ -804,6 +804,10 @@ class ExcelWriter(object):
     datetime_format : string, default None
         Format string for datetime objects written into Excel files
         (e.g. 'YYYY-MM-DD HH:MM:SS')
+    mode : {'w' or 'a'}, default 'w'
+        File mode to use (write or append).
+
+    .. versionadded:: 0.24.0
 
     Notes
     -----
@@ -897,7 +901,8 @@ def save(self):
         pass
 
     def __init__(self, path, engine=None,
-                 date_format=None, datetime_format=None, **engine_kwargs):
+                 date_format=None, datetime_format=None, mode='w',
+                 **engine_kwargs):
         # validate that this engine can handle the extension
         if isinstance(path, string_types):
             ext = os.path.splitext(path)[-1]
@@ -919,6 +924,8 @@ def __init__(self, path, engine=None,
         else:
             self.datetime_format = datetime_format
 
+        self.mode = mode
+
     def __fspath__(self):
         return _stringify_path(self.path)
 
@@ -993,23 +1000,27 @@ class _OpenpyxlWriter(ExcelWriter):
     engine = 'openpyxl'
     supported_extensions = ('.xlsx', '.xlsm')
 
-    def __init__(self, path, engine=None, **engine_kwargs):
+    def __init__(self, path, engine=None, mode='w', **engine_kwargs):
         # Use the openpyxl module as the Excel writer.
         from openpyxl.workbook import Workbook
 
-        super(_OpenpyxlWriter, self).__init__(path, **engine_kwargs)
+        super(_OpenpyxlWriter, self).__init__(path, mode=mode, **engine_kwargs)
 
-        # Create workbook object with default optimized_write=True.
-        self.book = Workbook()
+        if self.mode == 'a':  # Load from existing workbook
+            from openpyxl import load_workbook
+            book = load_workbook(self.path)
+            self.book = book
+        else:
+            # Create workbook object with default optimized_write=True.
+            self.book = Workbook()
 
-        # Openpyxl 1.6.1 adds a dummy sheet. We remove it.
-        if self.book.worksheets:
-            try:
-                self.book.remove(self.book.worksheets[0])
-            except AttributeError:
+            if self.book.worksheets:
+                try:
+                    self.book.remove(self.book.worksheets[0])
+                except AttributeError:
 
-                # compat
-                self.book.remove_sheet(self.book.worksheets[0])
+                    # compat - for openpyxl <= 2.4
+                    self.book.remove_sheet(self.book.worksheets[0])
 
     def save(self):
         """
@@ -1443,11 +1454,16 @@ class _XlwtWriter(ExcelWriter):
     engine = 'xlwt'
     supported_extensions = ('.xls',)
 
-    def __init__(self, path, engine=None, encoding=None, **engine_kwargs):
+    def __init__(self, path, engine=None, encoding=None, mode='w',
+                 **engine_kwargs):
         # Use the xlwt module as the Excel writer.
         import xlwt
         engine_kwargs['engine'] = engine
-        super(_XlwtWriter, self).__init__(path, **engine_kwargs)
+
+        if mode == 'a':
+            raise ValueError('Append mode is not supported with xlwt!')
+
+        super(_XlwtWriter, self).__init__(path, mode=mode, **engine_kwargs)
 
         if encoding is None:
             encoding = 'ascii'
@@ -1713,13 +1729,18 @@ class _XlsxWriter(ExcelWriter):
     supported_extensions = ('.xlsx',)
 
     def __init__(self, path, engine=None,
-                 date_format=None, datetime_format=None, **engine_kwargs):
+                 date_format=None, datetime_format=None, mode='w',
+                 **engine_kwargs):
         # Use the xlsxwriter module as the Excel writer.
         import xlsxwriter
 
+        if mode == 'a':
+            raise ValueError('Append mode is not supported with xlsxwriter!')
+
         super(_XlsxWriter, self).__init__(path, engine=engine,
                                           date_format=date_format,
                                           datetime_format=datetime_format,
+                                          mode=mode,
                                           **engine_kwargs)
 
         self.book = xlsxwriter.Workbook(path, **engine_kwargs)

pandas/tests/io/test_excel.py

@@ -2006,6 +2006,31 @@ def test_write_cells_merge_styled(self, merge_cells, ext, engine):
             assert xcell_b1.font == openpyxl_sty_merged
             assert xcell_a2.font == openpyxl_sty_merged
 
+    @pytest.mark.parametrize("mode,expected", [
+        ('w', ['baz']), ('a', ['foo', 'bar', 'baz'])])
+    def test_write_append_mode(self, merge_cells, ext, engine, mode, expected):
+        import openpyxl
+        df = DataFrame([1], columns=['baz'])
+
+        with ensure_clean(ext) as f:
+            wb = openpyxl.Workbook()
+            wb.worksheets[0].title = 'foo'
+            wb.worksheets[0]['A1'].value = 'foo'
+            wb.create_sheet('bar')
+            wb.worksheets[1]['A1'].value = 'bar'
+            wb.save(f)
+
+            writer = ExcelWriter(f, engine=engine, mode=mode)
+            df.to_excel(writer, sheet_name='baz', index=False)
+            writer.save()
+
+            wb2 = openpyxl.load_workbook(f)
+            result = [sheet.title for sheet in wb2.worksheets]
+            assert result == expected
+
+            for index, cell_value in enumerate(expected):
+                assert wb2.worksheets[index]['A1'].value == cell_value
+
 
 @td.skip_if_no('xlwt')
 @pytest.mark.parametrize("merge_cells,ext,engine", [
@@ -2060,6 +2085,13 @@ def test_to_excel_styleconverter(self, merge_cells, ext, engine):
         assert xlwt.Alignment.HORZ_CENTER == xls_style.alignment.horz
         assert xlwt.Alignment.VERT_TOP == xls_style.alignment.vert
 
+    def test_write_append_mode_raises(self, merge_cells, ext, engine):
+        msg = "Append mode is not supported with xlwt!"
+
+        with ensure_clean(ext) as f:
+            with tm.assert_raises_regex(ValueError, msg):
+                ExcelWriter(f, engine=engine, mode='a')
+
 
 @td.skip_if_no('xlsxwriter')
 @pytest.mark.parametrize("merge_cells,ext,engine", [
@@ -2111,6 +2143,13 @@ def test_column_format(self, merge_cells, ext, engine):
 
             assert read_num_format == num_format
 
+    def test_write_append_mode_raises(self, merge_cells, ext, engine):
+        msg = "Append mode is not supported with xlsxwriter!"
+
+        with ensure_clean(ext) as f:
+            with tm.assert_raises_regex(ValueError, msg):
+                ExcelWriter(f, engine=engine, mode='a')
+
 
 class TestExcelWriterEngineTests(object):