mfgd.md

lang.py.module.pandas.sorting.values

How to sort a dataframe based on column or row values

Synopsis

  df.sort_values(by=<[field(s)]>[, <extra-arguments>])

Overview

You can perform sorting on the entire dataframe with the sort_values() method. It takes only one mandatory argument than can be a single column (as a string) or a list of columns

Unlike sort_index this will perform sorting based on the values of a column or row

In case a list is provided, sorting will be performed in the order of the elements in the list

Arguments

• by:
  • Receives a list or a single string
  • The fields to sort by
  • Sorting order is based on the order of the elements in the list
• ascending:
  • Receives a boolean or a list of booleans
  • If True it sorts in ascending order (lower to higher)
  • If False it sorts in descending order (higher to lower)
• key
  • Receives a function or a lambda expression
  • Sort based on the results of mapping the passed function to each field in the column
• axis: Indicate to sort by rows (axis=0) or columns (axis=1)
  • You can also use columns passes as a string
• na_position: Whether you want rows with NaN to appear first
• kind: The type of sorting algorithm to use
  • This is ignored when sorting on more than one column
  • quicksort: Default sorting algorithm
  • mergesort: Stable sorting algorithm
  • heapsort: Heap style sorting algorithm

Cookbook

Sort a dataframe based on the values of a column

The simple way to sort a dataframe is by passing the name of the column to sort_values() this will by default sort the column in ascending order according to the values of that column

Sorting does not mutate your original dataframe, it returns a new sorted dataframe

  sortded_df = df.sort_values("Scores")

Sort a dataframe based on more than one column

You can sort a dataframe according to the values of more than one column by passing a list as the by argument. The order of the sorting will be based on the order of the elements in the list

  import pandas as pd

  df_exams = pd.read_csv('StudentsPerformance.csv')
  df_exams.sort_values(by=['math score', 'lang score'], ascending=[False, True])

# Output

# The dataframe sorted first by the 'math score' ind descendig order
# and then by the 'lang score' in ascending order

Sort based on processed value

You can perform sorting based on a different criteria by passing a function that processed the values of the column and return the criteria to sort on with the key argument

  import pandas as pd

  df_exams = pd.read_csv('StudentsPerformance.csv')
  df_exams.sort_values(by='race', ascending=True, key=lamda col: col.str.lower())

Change the sorting algorithm

You can change the sorting algorithm with the kind argument. It can take any of three merge algorithm

• quicksort: Default sorting algorithm
• mergesort: Stable sorting algorithm
  • It will maintain the original order of the records with the same key
  • Necessary if you plan to perform multiple sorts
• heapsort: Heap style sorting algorithm

This will be ignored if you are sorting on more than one column or label

  df.sort_values(
      by="cityId",
      ascending=False,
      kind="mergesort"
  )

Sort based on missing values

By default, when performing sorting, regardless of the sorting order, missing values will go at the end of the dataframe.

You can use the na_position='first' to put them at the begining of your dataframe

  df.sort_values(by='Notas', na_position='first')