Sort Dataframe by the counts of the column values

Question

I have a pandas dataframe (df) and I need to sort it based on the counts of the values of a column. The values of the column are string.

For example, the target column's values are orange, apple, banana, peach. The individual counts ( df['fruit'].value_counts() ) are:

• banana 2678
• peach 2250
• orange 1765
• apple 1691

The result I need the initial dataframe (with all the columns etc) sorted based on these counts. So, in the first 2678 rows the value in the fruit column should be banana etc

Answer 1

you can use map then sort_values for what you want.

df['sort_col'] = df['fruit'].map(df['fruit'].value_counts())
df = df.sort_values(by='sort_col', ascending=False).drop('sort_col', axis=1)