如何将DataFrame的D列与numpy ndarray比较并更新E列值？

Got it, let's solve this Pandas and numpy matching problem step by step!

Problem Statement #

Suppose we have this Pandas DataFrame df:

A   B   C   D          E
0  1   2   4   6  该列需更新
1 12  34   5  54           
2  4   8  12   4           
3  3   5   6   2           
4  5   7  11  27           

And a numpy ndarray npar with shape (4,1):

npar = np.array([[12], [6], [2], [27]])

We need to update the E column in df: if the value in column D exists anywhere in npar, set E to 1; otherwise, set it to 0.

Solution Code #

Here are two straightforward and efficient ways to get this done:

Method 1: Use isin() (Recommended) #

This is the most concise and performant approach. First, we flatten the 2D numpy array to a 1D array, then use Pandas' isin() method to check for matches, and convert the boolean results to integers (since True maps to 1 and False maps to 0):

import pandas as pd
import numpy as np

# Create the sample DataFrame
df = pd.DataFrame({
    'A': [1, 12, 4, 3, 5],
    'B': [2, 34, 8, 5, 7],
    'C': [4, 5, 12, 6, 11],
    'D': [6, 54, 4, 2, 27],
    'E': ['该列需更新', '', '', '', '']
})

# Create the sample numpy array
npar = np.array([[12], [6], [2], [27]])

# Update the E column
df['E'] = df['D'].isin(npar.flatten()).astype(int)

# Print the result to verify
print(df)

Method 2: Use apply() with any() #

If you prefer a more explicit row-wise check (good for understanding the logic), you can use apply() with a custom function that checks if each value in column D exists in the numpy array:

# Same setup for df and npar as above

def check_match(val):
    # Check if the value is present anywhere in the numpy array
    return 1 if (npar == val).any() else 0

# Apply the function to column D and update E
df['E'] = df['D'].apply(check_match)

# Print the result
print(df)

Final Output #

Both methods will produce the following updated DataFrame, where E is correctly set to 1 for matching values and 0 otherwise:

A   B   C   D  E
0  1   2   4   6  1
1 12  34   5  54  0
2  4   8  12   4  0
3  3   5   6   2  1
4  5   7  11  27  1

内容的提问来源于stack exchange，提问作者ASING