CODEWITHSUNDEEP
pythonpython-3.xpandasdataframedata-science
Ronak Thakkar
Ronak Thakkar

Reputation: 59

Insert list in pandas dataframe cell

I have a dictionary where each key has a list of values. Length of the list associated with each key is different. I want to convert the dictionary into a pandas dataframe with two columns 'Key' and 'Values'. Each row having one dictionary key in the 'Key' column and the list of values associated with it in 'Values' column. The dataframe will look as follows:

mapping_dict = {'A':['a', 'b', 'c', 'd'], 'B':['aa', 'bb', 'cc']}

df = 
    Key   Value
0   A     ['a', 'b', 'c', 'd']
1   B     ['aa', 'bb', 'cc']

I tried using the answer provided here by modifying it as per my use case. But it didn't output the required answer.

Upvotes: 2

Views: 19071

Answers (5)

BruceWayne74
BruceWayne74

Reputation: 41

You can do it simply by using the "at" setter from pandas:

df.at[0,'A'] = ['a', 'b', 'c', 'd']

It works if column dtype is "object".

Upvotes: 2

Alexis
Alexis

Reputation: 521

I had to insert a list at a specific location and the following solution worked for me.

For some reason pandas doesn't complain when returning a list/array in an apply function so:

df.loc[0,'A'] = df.loc[0,A].apply(lambda x: ['a', 'b', 'c', 'd'])

Upvotes: 2

Afshin Amiri
Afshin Amiri

Reputation: 3583

If you pass a list, pandas considers it as several rows. However, you can trick it by placing your list as the single element of an outer list as bellow:

import pandas as pd
mapping_dict = {'A':[['a', 'b', 'c', 'd']], 'B':[['aa', 'bb', 'cc']]}
df = pd.DataFrame(mapping_dict)
df

        A                 B
0   [a, b, c, d]    [aa, bb, cc]

Upvotes: 3

Keith
Keith

Reputation: 4914

I think you might have to update your dictionary beforehand then you can use from_dict. Update to make your dictionary to make it a list of list.

import pandas as pd
mapping_dict = {'A':['a', 'b', 'c', 'd'], 'B':['aa', 'bb', 'cc']}
updated_dict = {k: [v] for k, v in mapping_dict.items()}
df = pd.DataFrame.from_dict(updated_dict,orient='index')

If you want your exact formatting

df_formatted = df.reset_index()
df_formatted.columns = ['Key', 'Value']
print(df_formatted)

  Key         Value
0   B  [aa, bb, cc]
1   A  [a, b, c, d]

UPDATE

Bharath's answer is shorter but if you still want to use from_dict then you can take part of his method to do

df2 = pd.DataFrame.from_dict(list(mapping_dict.items()))
df2.columns = ['Key', 'Value']

Upvotes: 0

Bharath M Shetty
Bharath M Shetty

Reputation: 30605

Use pd.Series inside constructor, since dict values sizes are not equal, then set_axis to add column names i.e 

mapping_dict = {'A':['a', 'b', 'c', 'd'], 'B':['aa', 'bb', 'cc']}

df = pd.DataFrame(pd.Series(mapping_dict).reset_index()).set_axis(['Key','Value'],1,inplace=False)

  Key         Value
0   A  [a, b, c, d]
1   B  [aa, bb, cc]

Option 2 , convert the dict items to list then pass it to constructor: 

df = pd.DataFrame(list(mapping_dict.items()),columns=['Key','Value'])

Upvotes: 2

Related Questions