how to add a constant column to a dataframe without rows

I am trying to add a column with a constant value to a dataframe that does not have any rows. It appears this isn’t as easy as it would be if the rows were populated. How would one accomplish this?

df = pd.DataFrame(columns = ['a','b','c'])
df['foo'] = 'bar'

Should yield

 a   b   c   foo
0 NaN NaN NaN bar

instead it yields

  a   b   c 

Answer

You can use .loc specifying the row index and column label, as follows:

df.loc[0, 'foo'] = 'bar'

Result:

print(df)

     a    b    c  foo
0  NaN  NaN  NaN  bar

You can also use:

df['foo'] = ['bar']

Result:

print(df)

     a    b    c  foo
0  NaN  NaN  NaN  bar

If you have a bunch of a mix of empty and non-empty dataframes and you want to assign new column to it, you can try the following code:

df['foo'] = ['bar'] * (df.shape[0] if df.shape[0] else 1)

This will assign the constant with the same length (number of rows) for non-empty dataframes and will also assign one new row for empty dataframe with the constant value for the column.