如何为Pandas DataFrame连续0/1分组分配连续编号？

Hey there! The key here is to use pandas' vectorized operations to detect when a group changes and assign cumulative IDs to each consecutive group—no slow row-wise loops required.

Step-by-Step Breakdown: #

1. Spot group transitions: Compare each value in col1 with the previous value using shift(). A mismatch means we’ve started a new group.
2. Generate sequential group IDs: Take the cumulative sum of these transition indicators. Each time a change is detected, the sum increments, giving a unique ID for each continuous group.

Code Implementation: #

import pandas as pd

# Your sample data
d = {'col1': [1,1,1,0,0,1,1,0,0,0,1,1]}
df1 = pd.DataFrame(data=d)

# Add the group number column
df1['col2'] = (df1['col1'] != df1['col1'].shift()).cumsum()

print(df1)

Expected Output: #

col1  col2
0      1     1
1      1     1
2      1     1
3      0     2
4      0     2
5      1     3
6      1     3
7      0     4
8      0     4
9      0     4
10     1     5
11     1     5

Why This Works: #

• df1['col1'].shift() shifts all values down by one, letting us compare each row to the one before it.
• The boolean series df1['col1'] != df1['col1'].shift() returns True whenever the value changes (including the first row, since shift() returns NaN there).
• cumsum() treats True as 1 and False as 0, creating a running total that increments exactly when a new group starts.

This method is fully vectorized, so it’s blazingly fast even for large datasets—far more efficient than any loop-based approach.

内容的提问来源于stack exchange，提问作者Gev_2000