How to store a numpy arrays in a column of a Pandas dataframe?

Question

Is it possible to store arbitrary numpy arrays as the values of a single column in a dataframe of Pandas?

The arrays are all 2-dimensional, and I intend to use them to calculate values for other columns in the same dataframe.

To provide some context of what I'm trying to do here:

Each array is an adjacency matrix of some network, and for each network I want to calculate its various characteristics (e.g. density, centralities, clustering coefficient, etc) which are in fact other columns in the same dataframe.

may be it's me, but I think that question is not clear enough — roman
– roman, Commented Oct 24, 2013 at 19:11
FWIW I don't think your intention sits well with your request. While you can store arbitrary objects as values, you can't really do much with them in a vectorized fashion. — DSM
– DSM, Commented Oct 24, 2013 at 19:13
@DSM, the arrays are really just adjacency matrices of different graphs, and the other columns in the same dataframe are various network characteristics calculated based on each matrix. Do you suggest that I should decompose the matrix and store each row of the matrix in a separate column? — MLister
– MLister, Commented Oct 24, 2013 at 19:16
@RomanPekar, I've edited and provided more info in my question. — MLister
– MLister, Commented Oct 24, 2013 at 19:25

Back2Basics · Accepted Answer · 2015-12-28 03:30:52Z

16

Store them as elements as you would do for any other data:

import numpy as np
import pandas as pd
a = np.arange(10).reshape(2,5)
b = np.arange(10, 20).reshape(2,5)
pd.DataFrame({'foo':[42,51], 'arr':[a,b]})
Out[10]: 
                                            arr  foo
0            [[0, 1, 2, 3, 4], [5, 6, 7, 8, 9]]   42
1  [[10, 11, 12, 13, 14], [15, 16, 17, 18, 19]]   51

Note that what you try to do sounds more to use a Panel.

edited Dec 28, 2015 at 3:30

Back2Basics

7,8462 gold badges35 silver badges50 bronze badges

answered Oct 24, 2013 at 19:12

Zeugma

32.3k9 gold badges73 silver badges85 bronze badges

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

3 Comments

Seanny123 Over a year ago

Note, Panel is now deprecated.

soulmachine Over a year ago

This df can not be stored in PyTables format, df.to_hdf('test.hdf', key='xxx', format='table') will fail

yamini goel Over a year ago

@Boud by this way the numpy arrays get stored as a Series. Is there any way I can save it as numpy.ndarray in dataframe?

cc7768 · Accepted Answer · 2013-10-24 19:09:57Z

0

What do you mean store arbitrary numpy arrays as the values of a column in a dataframe of Pandas?

Something like this?

import numpy as np
import pandas as pd


x = np.random.randn(50, 25)
random_frame = pd.DataFrame(x)

This will store the array x in a DataFrame where the column names are 0, 1, 2, 3... Could you clarify? I think this is more a comment, but I don't know if I can comment yet.

answered Oct 24, 2013 at 19:09

cc7768

3254 silver badges10 bronze badges

2 Comments

MLister Over a year ago

i want to store arrays as values of a single column inside a dataframe, if possible.

cc7768 Over a year ago

I'm guessing that the 2 dimensional is n x m, and not nx1 right? I don't know if you can store an n x m array as a single column of a dataframe. Possible that I just haven't seen it. Would love to see how though.

Collectives™ on Stack Overflow

How to store a numpy arrays in a column of a Pandas dataframe?

2 Answers 2

3 Comments

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

3 Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related