sort_values() by multiple columns in Pandas - Step-by-Step Execution

Concept Flow - sort_values() by multiple columns

Start with DataFrame

↓

Choose columns to sort by

↓

Specify ascending or descending for each

↓

Call sort_values() with columns and order

↓

DataFrame sorted by first column

↓

Within ties, sorted by second column

↓

Return sorted DataFrame

We start with a DataFrame, pick columns to sort by, specify order for each, then call sort_values(). The DataFrame is sorted first by the first column, then by the second within ties.

Execution Sample

Pandas

import pandas as pd

df = pd.DataFrame({
  'A': [2, 1, 2, 1],
  'B': [1, 2, 2, 1]
})

sorted_df = df.sort_values(by=['A', 'B'], ascending=[True, False])

Sort the DataFrame first by column 'A' ascending, then by column 'B' descending within ties.

Execution Table

Step	DataFrame State	Action	Result
1	[{'A':2,'B':1},{'A':1,'B':2},{'A':2,'B':2},{'A':1,'B':1}]	Start with original DataFrame	DataFrame with 4 rows as given
2	Same as step 1	Choose columns to sort by: ['A', 'B']	Ready to sort by 'A' then 'B'
3	Same as step 1	Specify ascending=[True, False]	Sort 'A' ascending, 'B' descending
4	Same as step 1	Call df.sort_values(by=['A','B'], ascending=[True,False])	Sort rows by 'A' ascending and 'B' descending within ties
5	[{'A':1,'B':2},{'A':1,'B':1},{'A':2,'B':2},{'A':2,'B':1}]	Within ties of 'A', sort by 'B' descending	Rows with A=1 sorted by B descending: 2 then 1 Rows with A=2 sorted by B descending: 2 then 1
6	[{'A':1,'B':2},{'A':1,'B':1},{'A':2,'B':2},{'A':2,'B':1}]	Return sorted DataFrame	Final sorted DataFrame

💡 Sorting complete: DataFrame sorted by 'A' ascending and 'B' descending within ties

Variable Tracker

Variable	Start	After Step 4	After Step 5	Final
df	[{'A':2,'B':1},{'A':1,'B':2},{'A':2,'B':2},{'A':1,'B':1}]	[{'A':2,'B':1},{'A':1,'B':2},{'A':2,'B':2},{'A':1,'B':1}]	[{'A':2,'B':1},{'A':1,'B':2},{'A':2,'B':2},{'A':1,'B':1}]	[{'A':2,'B':1},{'A':1,'B':2},{'A':2,'B':2},{'A':1,'B':1}]
sorted_df	None	[{'A':1,'B':2},{'A':1,'B':1},{'A':2,'B':2},{'A':2,'B':1}]	[{'A':1,'B':2},{'A':1,'B':1},{'A':2,'B':2},{'A':2,'B':1}]	[{'A':1,'B':2},{'A':1,'B':1},{'A':2,'B':2},{'A':2,'B':1}]

Key Moments - 3 Insights

Why does the sorting first order by column 'A' before column 'B'?

Because in the execution_table step 4, sort_values() uses the 'by' list order. It sorts first by 'A', then breaks ties by sorting 'B' as shown in step 5.

How does specifying ascending=[True, False] affect the sorting?

As seen in step 3 and 5, 'A' is sorted ascending (small to large), while 'B' is sorted descending (large to small) within groups of equal 'A'.

Does the original DataFrame 'df' change after sorting?

No, as shown in variable_tracker, 'df' remains unchanged. The sorted DataFrame is returned as 'sorted_df'.

Visual Quiz - 3 Questions

Look at the execution_table step 5, what is the order of rows with 'A' equal to 2?

A [{'A':2,'B':2}, {'A':2,'B':1}]

B [{'A':2,'B':1}, {'A':2,'B':1}]

C [{'A':2,'B':1}, {'A':2,'B':2}]

D [{'A':2,'B':2}, {'A':2,'B':2}]

💡 Hint

Check the 'Within ties' sorting in step 5 of execution_table where 'B' is descending for 'A' = 2

At which step does the DataFrame get sorted by the first column 'A'?

A Step 2

B Step 4

C Step 5

D Step 6

💡 Hint

Look at execution_table step 4 where sort_values() is called and sorting by 'A' happens

If ascending=[False, True] was used instead, how would the order of rows with 'A' = 1 change?

A Rows with 'A' = 1 would not change order

B Rows with 'A' = 1 would be sorted by 'B' descending (2 then 1)

C Rows with 'A' = 1 would be sorted by 'B' ascending (1 then 2)

D Rows with 'A' = 1 would be sorted by 'A' descending

💡 Hint

Refer to how ascending list controls order per column in execution_table steps 3 and 5

Concept Snapshot

sort_values(by=[col1, col2], ascending=[bool1, bool2])
- Sorts DataFrame by multiple columns in order
- First sorts by col1, then col2 within ties
- ascending list controls order per column
- Returns new sorted DataFrame, original unchanged

Full Transcript

This visual execution shows how pandas sort_values() works when sorting by multiple columns. We start with a DataFrame with columns 'A' and 'B'. We choose to sort by 'A' first, then 'B'. We specify ascending order for 'A' as True and for 'B' as False. The sorting happens in steps: first by 'A' ascending, then within equal 'A' values by 'B' descending. The original DataFrame remains unchanged, and a new sorted DataFrame is returned. Key moments include understanding the order of sorting columns, how ascending list affects each column, and that the original data is not modified. The quizzes test understanding of sorting order and effects of changing ascending flags.