如何使用Pandas命令将df_score_2转换为df_score_1的全矩阵格式？

Got it, let's walk through how to turn your long-form df_score_2 into the square matrix structure of df_score_1. Here's a step-by-step breakdown with Pandas commands:

Step 1: Fix Column Names (if needed) #

First, make sure your df_score_2 has clear column names. If you loaded it from a text file without headers, assign meaningful names first:

import pandas as pd

# Assuming your df_score_2 looks like this after loading from the text file
df_score_2 = pd.DataFrame([
    ['A', 'B', 1],
    ['A', 'C', 1],
    ['A', 'D', 2],
    ['B', 'C', 5],
    ['B', 'D', 1]
])
# Assign descriptive column names
df_score_2.columns = ['row', 'col', 'score']

Step 2: Add Reverse Entries (for Symmetric Matrix) #

Since df_score_1 is a symmetric matrix (e.g., A→B = 1 and B→A = 1), we need to add the reverse pairs that are missing from df_score_2:

# Create reverse entries by swapping the row and col columns
reverse_entries = df_score_2.rename(columns={'row': 'col', 'col': 'row'})
# Combine original and reverse entries into one DataFrame
full_entries = pd.concat([df_score_2, reverse_entries], ignore_index=True)

Step 3: Pivot to Square Matrix #

Use Pandas' pivot method to reshape the long-form data into a wide matrix:

# Pivot the data: row becomes index, col becomes columns, score fills the cells
matrix_df = full_entries.pivot(index='row', columns='col', values='score')

Step 4: Fill Missing Values and Set Diagonal to 0 #

Fill any empty cells (where no score exists) with 0, and ensure the diagonal (same row/column) is 0 to match df_score_1:

# Replace NaN values with 0
matrix_df = matrix_df.fillna(0)
# Manually set diagonal elements to 0 (in case any slipped through)
for idx in matrix_df.index:
    matrix_df.loc[idx, idx] = 0

Step 5: Reindex to Match df_score_1's Order #

Finally, reorder the rows and columns to exactly match the sequence in df_score_1 (A, B, C, D):

matrix_df = matrix_df.reindex(index=['A', 'B', 'C', 'D'], columns=['A', 'B', 'C', 'D'])

Final Result #

Your matrix_df will now have the exact structure of df_score_1:

A  B  C  D
A  0  1  1  2
B  1  0  5  1
C  1  5  0  0
D  2  1  0  0

(Note: The values differ from your sample df_score_1 because df_score_2 uses different non-zero scores, but the matrix structure is identical.)

内容的提问来源于stack exchange，提问作者d..b