如何将零件编号列表转换为以宽/长为行列的二维网格（Pandas）

Let me break down where your current code is going off track, and show you a straightforward way to build the grid you need.

The Problem with Your Current Approach #

Your initial code tries to manually collect columns/rows and build a list of triples, but then the numpy array slicing step is misaligned with your data structure—your li contains individual part entries, not a pre-formatted table where the first row is column names and first column is row labels. That's why the DataFrame isn't turning out as expected. The extra data_dict/grid_data code also seems disconnected from your main goal here.

The Correct Solution Using Pandas Pivot #

Pandas has built-in tools (pivot or pivot_table) that are perfect for reshaping your part data into the grid layout you want. Here's a step-by-step implementation:

1. First, Let's Define Sample Data (to match your example)

# Simulate your part_list structure
class Part:
    def __init__(self, part_number, width, length):
        self.part_number = part_number
        self.width = width
        self.length = length

part_list = [
    Part("no1", "width1", "len1"),
    Part("no2", "width2", "len1"),
    Part("no3", "width1", "len2"),
    Part("no4", "width2", "len2"),
    Part("no5", "width3", "len2"),
    Part("no6", "width1", "len3"),
    Part("no7", "width2", "len3"),
    Part("no8", "width3", "len3"),
    Part("no9", "width4", "len3"),
]

2. Convert Your Part List to a Raw DataFrame

First, we'll translate your part_list into a simple Pandas DataFrame where each row represents one part:

import pandas as pd

# Create a raw DataFrame from your part list
df_raw = pd.DataFrame([
    {
        "width": part.width,
        "length": part.length,
        "part_number": part.part_number
    }
    for part in part_list
])

3. Reshape into the Width-Length Grid

Use pivot to rearrange the data so:

• Rows are your length values
• Columns are your width values
• Cells contain the corresponding part_number (empty if no part exists for that combination)

# Build the grid using pivot
df_grid = df_raw.pivot(
    index="length",   # Use length values as row labels
    columns="width",  # Use width values as column labels
    values="part_number"  # Fill cells with part numbers
)

# Optional: Sort rows and columns to match your example's order
df_grid = df_grid.reindex(sorted(df_grid.index))  # Sort length rows
df_grid = df_grid[sorted(df_grid.columns)]        # Sort width columns

# Replace NaN (empty cells) with blank strings if preferred
df_grid = df_grid.fillna("")

4. The Final Output

Running this will give you exactly the grid you want:

width  width1 width2 width3 width4
length                            
len1      no1    no2              
len2      no3    no4    no5       
len3      no6    no7    no8    no9

Key Notes #

• No need to manually collect columns/rows with loops—Pandas handles extracting unique values from the raw data automatically.
• The pivot method is ideal here because each (width, length) combination has at most one part number. If you ever have multiple parts per combination, use pivot_table instead with an aggregation function (like first()).

内容的提问来源于stack exchange，提问作者wmfox3