Get a row of data in pandas as a dict

user_651174 0
Viewed 3001

To get a row of data in pandas by index I can do:

df.loc[100].tolist()

Is there a way to get that row of data as a dict, other than doing:

dict(zip(
    df.columns.tolist(),
    df.loc[100], tolist()
))
pandas python
6 Answers

Try with to_dict

df.loc[1].to_dict()
user_79645270

You will run into a problem if you have columns with non-unique names.

Demo:

>>> df = pd.DataFrame([[1,2,3,4,5], [6,7,8,9,10]], columns=['A', 'B', 'A', 'C', 'B'])                                     
>>> df                                                                                                                 
   A  B  A  C   B
0  1  2  3  4   5
1  6  7  8  9  10
>>> df.loc[1].to_dict()                                                                                                
{'A': 8, 'B': 10, 'C': 9}

If this can happen in your dataframe, make the columns unique before creating the dict.

Here's an idea to do so:

>>> from itertools import count 
>>>                                                                                       
>>> col_isdupe = zip(df.columns, df.columns.duplicated(keep=False))                                                    
>>> counters = {c:count() for c, dupe in col_isdupe if dupe}                                                           
>>> df.columns = ['{}_{}'.format(c, next(counters[c])) if c in counters else c 
...:              for c in df.columns]                               
>>> df                                                                                                                 
   A_0  B_0  A_1  C  B_1
0    1    2    3  4    5
1    6    7    8  9   10
>>>                                                                                                                    
>>> df.loc[1].to_dict()                                                                                                
{'A_0': 6, 'A_1': 8, 'B_0': 7, 'B_1': 10, 'C': 9}
user_36200030

You can use items:

dict(df.loc[100].items())
user_39916490

df.loc[x] returns a mapping, a pd.Series, so you can just use the dict constructor directly:

dict(df.loc[100])

Or the to_dict helper method if you prefer...

This sort of raises the question, are you sure you need a dict at all?

user_50144550

Pandas DataFrame to List of Dictionaries

You can use: df.to_dict('records')

Ref: pandas docs.

Example:

Say you have a df:

>>> df
Out[1]: 
          alpha      beta  ...  log_10_rate       rate
0      2.530809  3.446069  ...     1.299609  19.934677
1      1.418243  1.861504  ...     1.059239  11.461440
...         ...       ...  ...          ...        ...
11241  2.462758 -0.890800  ...     1.532919  34.112962
11242  2.462758 -0.890800  ...     1.437005  27.353010
[11243 rows x 21 columns]

To get the ith row as a dict, I would do:

df.to_dict('records')[11242]
Out[2]: 
{'alpha': 2.462758375498395,
 'beta': -0.8908002057212157,
 'mmax': 90.14711749088858,
...
}
user_101777590

Say your dataframe is df and you want the row with index k, you can do:

list(df.iloc[k,:])

The result will be a list of all the values in row k.

user_69473490
Related