Remove duplicates in DF and convert to JSON obj in python

Question content

I have a df similar to the one below

name         series
=============================
a             a1
b             b1
a             a2
a             a1
b             b2

I need to convert the series into a list which should be assigned to each name like dictionary or json obj like below

{
   "a": ["a1", "a2"],
   "b": ["b1", "b2"]
}

So far I have tried using groupby but it just groups everything into a single dictionary

test = df.groupby("series")[["name"]].apply(lambda x: x)

The above code gives a df-like output

Series
Name
A     0   A1
      2   A2
      3   A1
B     1   B1
      4   B2

Any help is greatly appreciated

Thank you

Correct answer

Firstdrop_duplicates Make sure there is, thengroupby. agg as a list:

out = df.drop_duplicates().groupby('name')['series'].agg(list).to_dict()

Or dial unique:

out = df.groupby('name')['series'].agg(lambda x: x.unique().tolist()).to_dict()

Output: {'a': ['a1', 'a2'], 'b': ['b1', 'b2']}

If you have additional columns, make sure to keep only the columns of interest:

out = (df[['name', 'series']].drop_duplicates()
       .groupby('name')['series'].agg(list).to_dict()
      )

Sort the list:

out = (df.groupby('name')['series']
         .agg(lambda x: sorted(x.unique().tolist())).to_dict()
      )

Example:

# input
  Name Series
0    A     Z1
1    B     B1
2    A     A2
3    A     Z1
4    B     B2

# output
{'A': ['A2', 'Z1'], 'B': ['B1', 'B2']}

The above is the detailed content of Remove duplicates in DF and convert to JSON obj in python. For more information, please follow other related articles on the PHP Chinese website!