Adding a multi-level column to a single-level pandas dataframe

Question

I have a single-level pandas dataframe:

df = pd.DataFrame({"x":[0,0,0],"y":[0,0,0]})

which looks like this:

Now I want to add a multi-level column "z" (with two sub-columns "z1 and "z2") to this dataframe, so it'll look like this:

    x   y   z
            z1  z2
0   0   0   1   2
1   0   0   1   2
2   0   0   1   2

your life is probably easier if you just add columns df["z1"] = 1; df["z2"] = 2 — anon01
– anon01, Commented Mar 31, 2021 at 0:45
@anon01 Well in reality I need multiple multi-level columns in my dataframe, each with multiple sub-columns, and I often need to manipulate all sub-columns of a given parent column. If I don't use multi-level columns, then each time I need to do a list comprehension to choose all columns that start with a given string (e.g. all columns that start with a "z", i.e. "z1" and "z2"), so it'll a bit messy. So, if there is an easy way to add multi-level columns to single-level dataframes, that'll make my job easier I guess. — AAriam
– AAriam, Commented Mar 31, 2021 at 1:01

piterbarg · Accepted Answer · 2021-03-31 12:41:41Z

2

First we make the existing columns multi-index:

df.columns = pd.MultiIndex.from_arrays([df.columns,['']*len(df.columns)])

and then add new ones indexed by tuples

df[('z','z1')] = [1,1,1]
df[('z','z2')] = [2,2,2]
df

to get

    x   y   z
            z1  z2
0   0   0   1   2
1   0   0   1   2
2   0   0   1   2

answered Mar 31, 2021 at 12:41

piterbarg

8,2292 gold badges9 silver badges22 bronze badges

Sign up to request clarification or add additional context in comments.

1 Answer 1