LXML : Populating Pandas DataFrame

Question

I am running the following code and it works perfectly to populate the DataFrame with one line of information. I will be parsing more information and appending to the DataFrame. I do not want to run:

d.setdefault('Stat_Name', [])

for each statistic that I want to parse and add. Is there a more efficient way to do this?

url = 'http://finance.yahoo.com/q/is?s=MMM+Income+Statement&annual'
tree = html.parse(url)
years = ['year1', 'year2', 'year3']
d = {}
d.setdefault('Total Revenue', [])
for i in range(2,5):
    contentB = tree.xpath("//table[@class='yfnc_tabledata1']/tr[1]/td/table/tr[3]/td[{i}]".format(i=i))[0].text_content().strip()
    d['Total Revenue'].append(int(contentB.replace(',', '')))
df = pd.DataFrame(columns=years, index=d.keys())
for eachItem in d.values():
    pass
df.loc['Total Revenue'] = eachItem
print(df)