How to assign dict type value to one element of a pandas dataframe

How to assign dict type value to one element of a pandas dataframe.
Pandas version is 1.5.3; python version is 3.9

for example:

import pandas as pd
test1_pd = pd.DataFrame([[1,2,3]], columns=['a','b','c'])
test2_pd = pd.DataFrame([[1,2,3]], columns=['a','b','c'])
test1_pd.loc[0]['a'] = {'x':1,'y':2}
test2_pd.loc[0]['a'] = 9

the result is:

test1_pd
Out[116]: 
   a  b  c
0  1  2  3

test2_pd
Out[117]: 
   a  b  c
0  9  2  3

Why test1_pd['a'][0] is not replaced by dict value {'x':1,'y':2}?
But test2_pd['a'][0] is replaced by 9 successfully.

Even though @Vvvvvv added the link to the answer, would like to add some interesting insights to the topic.

In your case, when you try to assign the dictionary, the default nature of pandas will try to look for the key 'a' inside the dictionary and assign the value of 'a' to the column 'a' in the dataframe.

test1_pd.loc[0]['a'] = {'a':4,'b':5}

Output:-

|a|b|c|<br>
|4|2|3|

(Apologies for displaying the output like this), The columns that are to be indexed ('a') in your case will be looked inside the dictionary, and hence no value changes when you have the keys as 'x' and 'y' because these variables are not the column names in your dataframe. Another example is

test1_pd.loc[0][['a','c']] = {'a':4,'c':5}

Output:-

|a|b|c|<br>
|4|2|5|

A quick fix to your problem will be -

test1_pd.loc[0, 'a'] = [{'x':1,'y':2}]

presenting the values in the form of a list where number of elements in the list will be equal to the number of values in the column.