Summing Multiple Columns With Multiindex Columns

I have a dataframe that is created from a pivot table, and looks similar to this: import pandas as pd d = {('company1', 'False Negative'): {'April- 2012': 112.0, 'April- 2013': 370

Solution 1:

You can calculate this sum by specifying the level (you want to sum along the first level (level 0), so collapsing the second level):

In [29]:df.sum(axis=1,level=0)Out[29]:company1company2April-2012        112112April-2013       1054      1054April-2014        573573August-2012       431431August-2013       496496August-2014       724724

If you want them to add to the original dataframe, as in your example above, you can add a level in the columns and concat:

sums = df.sum(level=0, axis=1)
sums.columns = pd.MultiIndex.from_product([sums.columns, ['SUM']])
df = pd.concat([df, sums], axis=1)