Merge pull request #5446 from jtratner/default-imports-for-excel-writers

PERF/ENH: Use xlsxwriter by default (and then fall back on openpyxl).

Author: jtratner

doc/source/io.rst

@@ -1825,8 +1825,10 @@ written.  For example:
 
    df.to_excel('path_to_file.xlsx', sheet_name='Sheet1')
 
-Files with a ``.xls`` extension will be written using ``xlwt`` and those with
-a ``.xlsx`` extension will be written using ``openpyxl``.
+Files with a ``.xls`` extension will be written using ``xlwt`` and those with a
+``.xlsx`` extension will be written using ``xlsxwriter`` (if available) or
+``openpyxl``.
+
 The Panel class also has a ``to_excel`` instance method,
 which writes each DataFrame in the Panel to a separate sheet.
 
@@ -1858,14 +1860,19 @@ Excel writer engines
 1. the ``engine`` keyword argument
 2. the filename extension (via the default specified in config options)
 
-By default, ``pandas`` uses  `openpyxl <http://packages.python.org/openpyxl/>`__
-for ``.xlsx`` and ``.xlsm`` files and `xlwt <http://www.python-excel.org/>`__
-for ``.xls`` files.  If you have multiple engines installed, you can set the
-default engine through :ref:`setting the config options <basics.working_with_options>`
-``io.excel.xlsx.writer`` and ``io.excel.xls.writer``.
+By default, ``pandas`` uses the `XlsxWriter`_  for ``.xlsx`` and `openpyxl`_
+for ``.xlsm`` files and `xlwt`_ for ``.xls`` files.  If you have multiple
+engines installed, you can set the default engine through :ref:`setting the
+config options <basics.working_with_options>` ``io.excel.xlsx.writer`` and
+``io.excel.xls.writer``. pandas will fall back on `openpyxl`_ for ``.xlsx``
+files if `Xlsxwriter`_ is not available.
+
+.. _XlsxWriter: http://xlsxwriter.readthedocs.org
+.. _openpyxl: http://packages.python.org/openpyxl/
+.. _xlwt: http://www.python-excel.org
 
-For example if the `XlsxWriter <http://xlsxwriter.readthedocs.org>`__
-module is installed you can use it as a xlsx writer engine as follows:
+To specify which writer you want to use, you can pass an engine keyword
+argument to ``to_excel`` and to ``ExcelWriter``.
 
 .. code-block:: python
 

pandas/core/config_init.py

@@ -300,8 +300,7 @@ def use_inf_as_null_cb(key):
 with cf.config_prefix('io.excel'):
     # going forward, will be additional writers
     for ext, options in [('xls', ['xlwt']),
-                         ('xlsm', ['openpyxl']),
-                         ('xlsx', ['openpyxl'])]:
+                         ('xlsm', ['openpyxl'])]:
         default = options.pop(0)
         if options:
             options = " " + ", ".join(options)
@@ -310,3 +309,17 @@ def use_inf_as_null_cb(key):
         doc = writer_engine_doc.format(ext=ext, default=default,
                                        others=options)
         cf.register_option(ext + '.writer', default, doc, validator=str)
+    def _register_xlsx(engine, other):
+        cf.register_option('xlsx.writer', engine,
+                           writer_engine_doc.format(ext='xlsx',
+                                                    default=engine,
+                                                    others=", '%s'" % other),
+                          validator=str)
+
+    try:
+        # better memory footprint
+        import xlsxwriter
+        _register_xlsx('xlsxwriter', 'openpyxl')
+    except ImportError:
+        # fallback
+        _register_xlsx('openpyxl', 'xlsxwriter')

pandas/io/tests/test_excel.py

@@ -14,7 +14,7 @@
 from pandas.io.parsers import read_csv
 from pandas.io.excel import (
     ExcelFile, ExcelWriter, read_excel, _XlwtWriter, _OpenpyxlWriter,
-    register_writer
+    register_writer, _XlsxWriter
 )
 from pandas.util.testing import ensure_clean
 from pandas.core.config import set_option, get_option
@@ -1040,9 +1040,14 @@ def test_ExcelWriter_dispatch(self):
         with tm.assertRaisesRegexp(ValueError, 'No engine'):
             ExcelWriter('nothing')
 
-        _skip_if_no_openpyxl()
+        try:
+            import xlsxwriter
+            writer_klass = _XlsxWriter
+        except ImportError:
+            _skip_if_no_openpyxl()
+            writer_klass = _OpenpyxlWriter
         writer = ExcelWriter('apple.xlsx')
-        tm.assert_isinstance(writer, _OpenpyxlWriter)
+        tm.assert_isinstance(writer, writer_klass)
 
         _skip_if_no_xlwt()
         writer = ExcelWriter('apple.xls')