Pandas Dataframe Examples: Create and Append data

Pandas Dataframe Examples: Create and Append data

Last updated:
Table of Contents

Examples using Pandas 2.x

There are many ways to build and initialize a pandas DataFrame. Here are some of the most common ones:

All examples can be found on this notebook

Create from lists

Where each list represents one column.

import pandas as pd

names = ['john','mary','peter','gary','anne']
ages = [33,22,45,23,12]

df = pd.DataFrame({
    'names':names,
    'ages':ages
})
df

dataframe-built-using-lists Probably the most straightforward
way to build dataframes

Create from dicts

To create a dataframe from a list of dicts use pd.DataFrame.from_records().

import pandas as pd

data_dicts = [
    {'name':"john","gender":'male','age':45},
    {'name':"mary", 'gender':"female",'age':19},
    {'name':"peter",'gender':'male', 'age':34}
]

df = pd.DataFrame.from_records(data_dicts)
df

pandas-dataframe-create-from-list-of-dicts Since we didn't specify dtypes, they are automatically inferred from the data.

Create from dict

To create a dataframe from a single dict using keys the index use pd.DataFrame.from_dict(my_dict, orient='index')

import pandas as pd

d = {"alice": 12, "bob": 20, "charlie": 33}

pd.DataFrame.from_dict(d, orient='index')

source-dict SOURCE DICT: just a simple dict
  
dataframe-built-using-dict-keys-as-index You can set the name of the column
if you want, passing columns=['age'] to from_dict

Create empty Dataframe, append rows

Use append() with ignore_index=True.

import pandas as pd

# if you wish, you can set column names and dtypes here
df = pd.DataFrame()

# must reassign since the append method does not work in place
df = df.append({'col_a':5,'col_b':10}, ignore_index=True)
df = df.append({'col_a':1,'col_b':100}, ignore_index=True)
df = df.append({'col_a':32,'col_b':999}, ignore_index=True)

df

dataframe-create-empty-and-append Since ignore_index is set, indices will start at 0

Create with dtypes

As of version 2.1 this is not possible!

Only one dtype can be passed, which only works if all columns are of that type!

A workaround is to call astype() on every column after initialization.