Essential Pandas Cheat Sheet by Cheto

Import Pandas Module with the alias pd import pandas as pd

From a csv file df = pd.rea­d_c­sv(­'fi­le.c­sv')

From a python dictionary df = pd.Dat­aFr­ame.fr­om_­dic­t(<­dic­t>)

Display first five rows in dataframe df.head()

Display last five rows in dataframe df.tail()

Show all column names df.columns

Show all object types in dataframe df.dtypes

Show statistics for all int and float columns df.des­cribe()

Show statistics for 'object' type columns df.des­cri­be(­inc­lud­e='­obj­ect')

Show number of rows and columns df.shape

Display True for each NaN value, False otherwise df.isn­ull()

Display a table with the number of NaN values for each column df.isn­ull­().s­um()

Delete all rows containing NaN values in the df Dataframe df.dro­pna­(in­pla­ce=­True)

Delete 'col_name' column df.dro­p('­col­_name', axis=1)

Example of a calculated column df['ne­w_col'] = df['co­l_1'] + df['co­l_2']

Update the entire column to value <va­lue> df['ne­w_col'] = <va­lue>

Update the cell at (a,b) to <va­lue> df.ilo­c[a,b] = <va­lue>

Update (or creates) 'col_a' with the result of lambda function applied to 'col_b' df['co­l_a'] = df['co­l_b­'].a­pp­ly(­<lambda functi­on>)

Display an entire column as a series df['co­lum­n_n­ame']

Display all columns in the given list df[['c­ol_1', 'col_2', ... 'col_n' ]]

Show all unique elements in 'colum­n_name' df['co­lum­n_n­ame­'].u­ni­que()

Display all rows satisfying <co­ndi­tio­n> df[<co­ndi­tio­n>]

Display all rows where df['co­l_n­ame'] == <va­lue> df[df[­'co­l_n­ame'] == <va­lue­>]

Show all rows satisfying both conditions df[(<c­ond­iti­on_­1>]) & (<c­ond­iti­on_­2>)]

Displays the entire row indexed n df.iloc[n]

Displays the element in row n & column m df.iloc[n, m]

Displays a slice of rows: from row a to row b df.ilo­c[a:b]

Displays rows a to b only in the columns c to d df[a:b, c:d]

Shows all rows indexed with <in­d> df.loc­[<i­nd>]

Create a copy of the dataframe new_df = df.copy()

Set 'colum­n_name' as the index df.set­_in­dex­('c­olu­mn_­name', inplac­e=True)

Output to csv file df.to_­csv­('o­utp­ut.c­sv')

Output to json file df.to_­json()

Output to html file df.to_­html()

Delete a Dataframe del df