Excel parser cannot change *keep_default_na* value

[davidshinn]

The parse method in the ExcelFile class does not pass the argument keep_default_na to the TextParser from pandas.io.parsers. There is no clean way to override the default na values when parsing excel files with the current code.

My specific problem involves "NA" as a typical value for "North America" in excel files.

[jtratner]

@davidshinn I think #4139 should fix this - just changed parse and _parse_excel to pass keyword arguments through

[davidshinn]

@jtratner This fixes the problem perfectly and other potential kwds missing from upstream functions. I'm extremely impressed with the quick commits. On another note, I noticed that because this doesn't specifically add the keyword argument, the feature is not as visible through ipython tab completion and introspection. However from my perspective, the main issue is closed. Maybe I'll tackle those niceties myself. Thanks.

[jtratner]

@davidshinn Especially now that ipython offers tab completion on keyword arguments, I agree. The sole issue with doing this is that it adds a little overhead to making sure that the default arguments of TextParser and ExcelFile.parse match up later on.