Pandas - Dump dataframe columns into dictionary

Question

I want to dump a two-columns dataframe into a dictionary like this:

My dataframe:

df['col1'] = pd.Series(['id_1','id_2','id_3','id_4'])
df['col2'] = pd.Series(['Name1','Name2','Name3','Name4'])

Desired dictionary

dict = {'id_1':'Name1','id_2':'Name2','id_3':'Name3',...}

What would be a nice way of doing this without looping through the rows?

Answer 1

Use df.set_index() to set the dictionary keys and take only the column you want as values and call series.to_dict() on it:

my_dict=df.set_index('col1')['col2'].to_dict()

Or you can use zip to zip 2 columns and call the dict function on it.

my_dict=dict(zip(df.col1,df.col2))

---

Output:

{'id_1': 'Name1', 'id_2': 'Name2', 'id_3': 'Name3', 'id_4': 'Name4'}

PS Don't use dict as a dictionary name since it is a python builtin func.

Answer 2

Another way is calling dict directly on numpy array of df .

If your df has exact 2 columns:

dict(df.values)

If your df has more than 2 columns, just slicing 2 columns you want to convert and call dict on them:

dict(df[['col1', 'col2']].values)    

Output:

{'id_1': 'Name1', 'id_2': 'Name2', 'id_3': 'Name3', 'id_4': 'Name4'}