How to Convert Pandas DataFrame to Dictionary   

In this short guide, I'll show you how to convert Pandas DataFrame to dictionary. You can also find how to use Pandas method - to_dict(). So at the end you will get from DataFrame to Python dict:

to:

{'day': {0: 1, 1: 2, 2: 3, 3: 4, 4: 5, 5: 6},
'numeric': {0: 1, 1: 2, 2: 3, 3: 4, 4: 5, 5: 6}}

or any other dict-like format. We will also cover the following examples:

• dict (default) : dict like {column -> {index -> value}}
• list : dict like {column -> [values]}
• series : dict like {column -> Series(values)}
• split : dict like - {'index' -> [index], 'columns' -> [columns], 'data' -> [values]}
• tight : dict like
  {'index' -> [index], 'columns' -> [columns], 'data' -> [values], 'index_names' -> [index.names], 'column_names' -> [column.names]}
• records : list like - [{column -> value}, ... , {column -> value}]
• index : dict like {index -> {column -> value}}

To start, here is the syntax that you may apply in order to convert DataFrame to dict:

df.to_dict()

In the next section, I'll review the steps to apply the above syntax in practice.

Step 1: Create a DataFrame

Lets create a DataFrame which has a two columns:

import pandas as pd

data={'day': [1, 2, 3, 4, 5, 6],
     'numeric': [1, 2, 3, 4, 5, 6]}

df = pd.DataFrame(data)
df

result:

Step 2: DataFrame to dict - {column -> {index -> value}}

In order to extract DataFrame as Python dictionary we need just this line:

df.to_dict()

result:

{'day': {0: 1, 1: 2, 2: 3, 3: 4, 4: 5, 5: 6},
'numeric': {0: 1, 1: 2, 2: 3, 3: 4, 4: 5, 5: 6}}

By default method to_dict() use as parameter - orient='list' and will produce dict form of:

{column -> {index -> value}}

Step 3: DataFrame to dict - list - {column -> [values]}

What if you like to get a dictionary only with the values? In this case we will use orient='list' in order to exclude index from the output dictionary:

df.to_dict(orient='list')

{'day': [1, 2, 3, 4, 5, 6], 'numeric': [1, 2, 3, 4, 5, 6]}

Note: have in mind that there is a difference between Python dict and JSON format.

Step 3: split - {'index' -> [index], 'columns' -> [columns], 'data' -> [values]}

Another form of dictionary which includes:

• index
• columns
• data

as separate lists can be extracted by:

df.to_dict(orient='split')

{'index': [0, 1, 2, 3, 4, 5],
 'columns': ['day', 'numeric'],
 'data': [[1, 1], [2, 2], [3, 3], [4, 4], [5, 5], [6, 6]]}

and then:

df['yyyy'].astype(str) + '-'+ df['mm'].astype(str)

Step 4: Convert DataFrame to dict - tight format

If you like to get a tight format of DataFrame as dict we can use the parameter orient='tight'.

This option includes the label names - index and column names in the output dict:

df.to_dict(orient='tight')

to get:

{'index': [0, 1, 2, 3, 4, 5],
'columns': ['day', 'numeric'],
'data': [[1, 1], [2, 2], [3, 3], [4, 4], [5, 5], [6, 6]],
'index_names': [None],
'column_names': [None]}

Step 5: Convert DataFrame to dict - records

If we like to extract only the values and the column labels as dictionary we can do it by - orient='records':

df.to_dict(orient='records')

We will get the list of dicts from the original DataFrame:

[{'day': 1, 'numeric': 1},
 {'day': 2, 'numeric': 2},
 {'day': 3, 'numeric': 3},
 {'day': 4, 'numeric': 4},
 {'day': 5, 'numeric': 5},
 {'day': 6, 'numeric': 6}]

Step 6: Convert DataFrame to dict - index

Finally we can convert DataFrame to dictionary with index by:

df.to_dict(orient='index')

result:

{0: {'day': 1, 'numeric': 1},
 1: {'day': 2, 'numeric': 2},
 2: {'day': 3, 'numeric': 3},
 3: {'day': 4, 'numeric': 4},
 4: {'day': 5, 'numeric': 5},
 5: {'day': 6, 'numeric': 6}}

More information about method - to_dict