Initialize dataframe with a constant

Question

Initializing dataframe with a constant value does not work,

pd.DataFrame(0, index=[1,2,3]) # doesnt work! # OR pd.DataFrame(0) # doesnt work!

whereas I observe that

(1) Initializing Series with a constant value works

pd.Series(0, index=[1,2,3]) # Works fine!

(2) Initializing DataFrame with None works

pd.DataFrame(None, index=[1,2,3]) # Works fine!

(3) Initializing DataFrame when index and columns are not provided works

pd.DataFrame([1, 2, 3]) # Works fine! pd.DataFrame([0]) # Works fine!

Does anyone know why?

I am keen to know more from a design consideration, rather than an answer like "if you check the pandas code, you'd see that it fails one of the checks where the data dimension is expected to be >1.. blah blah".

I think it should just work intuitively (considering that pandas is intelligent in assigning a default col and index when not provided, and also guesses the dimensions from the data given).

May be there is some reason for this behaviour but can't figure.

piRSquared · Accepted Answer · 2017-07-26 05:55:34Z

A pd.DataFrame is 2-dimensional. When you specify

pd.DataFrame(0, index=[1, 2, 3])

You are telling the constructor to assign 0 to every row with indices 1, 2, and 3. But what are the columns? You didn't define any columns.

You can do two things

Option 1
specify your columns

pd.DataFrame(0, index=[1, 2, 3], columns=['x', 'y']) x y 1 0 0 2 0 0 3 0 0

Option 2
pass a list of values

pd.DataFrame([[0]], index=[1, 2, 3]) 0 1 0 2 0 3 0

Collectives™ on Stack Overflow

Initialize dataframe with a constant

1 Answer 1

Comments

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Related