Convert Pandas Object to Array

Question

I have a column in dataframe which is of the type object. I want to convert it to an array. ex:

'["399866273","12444645","162497334"]'

I am aware of converting it to string and splitting it based on ,. But I want to know if there is a better way to do this.

If you mean a numpy.ndarray then you just want my_df['column'].values, if you want a list, then you just want my_df['column'].to_list() — juanpa.arrivillaga
– juanpa.arrivillaga, Commented Jul 30, 2020 at 21:23
I can't do this, it would say 'str' object has no attribute 'values' @juanpa.arrivillaga — John Constantine
– John Constantine, Commented Jul 30, 2020 at 21:27
Then you don't have a dataframe you have str. How did you produce this string? You need to provide a minimal reproducible example. — juanpa.arrivillaga
– juanpa.arrivillaga, Commented Jul 30, 2020 at 21:30
Try dir(SomeObject) for the object you are getting. It may help determine what you're dealing with. — Mike67
– Mike67, Commented Jul 30, 2020 at 21:31

skullgoblet1089 · Accepted Answer · 2020-07-30 22:05:32Z

1

import json
import numpy as np

def toarray(s):
    x = json.loads(s)
    arr = np.array(x)
    return arr

# Test
s = '["399866273","12444645","162497334"]'
asset type(toarray(s)) == np.ndarray
assert all(toarray(s) == np.array(["399866273", "12444645", "162497334"]))

# Apply to df
colname = ...  # your col name
df['as_array'] = df[colname].apply(toarray)

answered Jul 30, 2020 at 22:05

skullgoblet1089

6145 silver badges12 bronze badges

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

Comments

iadi7ya · Accepted Answer · 2020-07-30 22:27:43Z

You can try something like this:

df = pd.DataFrame(['["399866273","12444645","162497334"]'], columns = ['Column_ObjectType'])

df.dtypes
#output
Column_ObjectType    object
dtype: object

df[['Value1','Value2','Value3']] = df.Column_ObjectType.str.split(",", expand=True)
df.head()
#output
                      Column_ObjectType        Value1      Value2        Value3
0  ["399866273","12444645","162497334"]  ["399866273"  "12444645"  "162497334"]

df['Value1'] = df.Value1.apply(lambda x: x[2:-1])
df['Value2'] = df.Value1.apply(lambda x: x[1:-1])
df['Value3'] = df.Value1.apply(lambda x: x[1:-1])
df.head()
#output
                  Column_ObjectType     Value1    Value2    Value3
0  ["399866273","12444645","162497334"]  399866273  39986627  39986627

atszet · Accepted Answer · 2020-07-30 22:01:28Z

0

If you mean a list you can just use eval()

answered Jul 30, 2020 at 22:01

atszet

1

Collectives™ on Stack Overflow

Convert Pandas Object to Array

3 Answers 3

Comments

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related