How to insert a row into a DataFrame which has a tuple as an index?

Question

My DataFrame index is a tuple and I want to insert a row.

I thinked the code below would work:

df.loc[('first', 'second')] = (1, 2, 3)

But it doesn't do what I expect, instead it adds an idex 'first' and a column 'second' to a DataFrame.

You could do `df.loc[[('first', 'second')]]`, but check the dupe, which suggestes not to use tuples as index — yatu, May 17 '19 at 12:24

score 0 · Accepted Answer · answered May 17 '19 at 12:22

First convert index to MultiIndex by MultiIndex.from_tuples, then also also add list for select columns:

df = pd.DataFrame({
         'a':[1,3,5],
         'b':[5,4,5],
         'c':[10,11,12],
}, index= [('a','b'),('a','c'),('a','d')])

print (df)
        a  b   c
(a, b)  1  5  10
(a, c)  3  4  11
(a, d)  5  5  12

df.index = pd.MultiIndex.from_tuples(df.index)
df.loc[('first', 'second'), ['a','b','c']] = (1,2,3)
print (df)
                a    b     c
a     b       1.0  5.0  10.0
      c       3.0  4.0  11.0
      d       5.0  5.0  12.0
first second  1.0  2.0   3.0

How to insert a row into a DataFrame which has a tuple as an index?

1 Answers1