read_excel does not accept a URL

[jseabold]

http://nbviewer.ipython.org/urls/umich.box.com/shared/static/oh717lkxczhseep71lao.ipynb

[asobrien]

What you want is to directly open an xls file by directly specifiy the URL in pd.read_excel:

pd.read_excel("http://www.eia.gov/dnav/pet/xls/PET_PRI_ALLMG_A_EPM0_PTC_DPGAL_M.xls", "Data 1", skiprows=2)

but currently this returns an IOError:

IOError: [Errno 2] No such file or directory: 'http://www.eia.gov/dnav/pet/xls/PET_PRI_ALLMG_A_EPM0_PTC_DPGAL_M.xls'

Getting around this currently requires explicit calls to urllib2 and StringIO to read and generate a buffer which can then be passed to pd.read_csv:

data_url = "http://www.eia.gov/dnav/pet/xls/PET_PRI_ALLMG_A_EPM0_PTC_DPGAL_M.xls"
xld = urllib2.urlopen(data_url).read()
xlds = StringIO.StringIO(xld)
data = pd.read_excel(xlds, "Data 1", skiprows=2)

[asobrien]

I'm interested in implementing a method to directly read from a URL in pd.read_excel. My initial thought is to utilize a is_url boolean keyword argument to explicitly indicate that the io string is indeed a URL. The method could then simply be implemented using urllib2 and StringIO, as above.

Any thoughts?

[jreback]

see #7531