Pandas >> How to Create DataFrame in Pandas

Table of Contents

In this article, we will talk about different ways of creating DataFrame in Pandas.

Give dict data in pandas.DataFrame constructor directly

import pandas as pd

df = pd.DataFrame({
    "name": ["Kevin", "Jack", "Mary"],
    "score": [80, 90, 95]
})

Give dict data outside pandas.DataFrame constructor

We can difine a dict data outside pandas.DataFrame constructor and specify the data by using data option.

import pandas as pd

d = {
    "name": ["Kevin", "Jack", "Mary"],
    "score": [80, 90, 95]
}

df = pd.DataFrame(data=d)

Give a list of list data outside pandas.DataFrame constructor

We can difine a list of list data outside pandas.DataFrame constructor and specify the data by using data and columns option.

import pandas as pd

d = [["Kevin", 80], ["Jack", 90], ["Mary", 95]]

df2 = pd.DataFrame(data=d, columns=["name", "score"])

Result

Use pandas.DataFrame.from_records method

We can use pandas.DataFrame.from_records to creates a DataFrame object from a structured ndarray, sequence of tuples or dicts, or DataFrame.

• structured ndarray

We can define a numpy array and use pandas.DataFrame.from_records to create DataFrame.

import pandas as pd

data = np.array([('N001', 1), ('N002', 2), ('N003', 3)],
                dtype=[('id', 'U1'), ('num', 'i4')])

df = pd.DataFrame.from_records(data)

• a list of dicts

We can define a list of dicts and use pandas.DataFrame.from_records to create DataFrame.

import pandas as pd

data = [{'id': 1, 'name': 'a'},
        {'id': 2, 'name': 'b'},
        {'id': 3, 'name': 'c'},
        {'id': 4, 'name': 'd'}]

df = pd.DataFrame.from_records(data)

• a list of tuples

We can define a list of tuples and use pandas.DataFrame.from_records to create DataFrame.

import pandas as pd

data = [(1, 'a'), (2, 'b'), (3, 'c'), (4, 'd')]

df = pd.DataFrame.from_records(data, columns=["id", "name"])

Use pandas.DataFrame.from_dict method

We can use pandas.DataFrame.from_dict to creates a DataFrame object from dictionary by columns or by index allowing dtype specification.

• the keys of dict become the DataFrame columns

We can define a dict and use pandas.DataFrame.from_dict to create DataFrame, and its keys become DataFrame columns.

import pandas as pd

data = {'id': [1, 2, 3, 4], 'name': ['a', 'b', 'c', 'd']}

df = pd.DataFrame.from_dict(data)

Result

• the keys of dict become the DataFrame rows

If we specify orient='index', the keys will become rows.

import pandas as pd

data = {'id': [1, 2, 3, 4], 'name': ['a', 'b', 'c', 'd']}

df = pd.DataFrame.from_dict(data, orient='index')

Result

• specify columns

When using the index orientation, the column names can be specified manually.

We can define a list of tuples and use pandas.DataFrame.from_records to create DataFrame.

import pandas as pd

data = {'id': [1, 2, 3, 4], 'name': ['a', 'b', 'c', 'd']}

df = pd.DataFrame.from_dict(data, orient='index', columns=['A', 'B', 'C', 'D'])

Result