Python生成的3D Plot未包含全部列数据，求解决方案

Hey John, let's figure out why your 3D plot isn't showing all 109 cases. Since your CSV handling and statistics code works fine, the problem is almost certainly in how you're preparing or passing data to your plotting library (I’m assuming you’re using matplotlib’s 3D toolkit, since that’s the standard for this in Python). Here are the most common fixes to try:

1. Confirm You’re Loading All 109 Rows First #

First, double-check that the data you’re feeding into the plot actually has all 109 rows. Add a quick print statement right after loading your data from dump2.csv:

# After loading your x, y, z data arrays
print(f"Total rows loaded: {len(x_data)}")

If this doesn’t output 109, you might have:

• Accidentally skipped rows (like a misplaced next(reader) call if your CSV has a header, but you’re skipping it multiple times)
• Filtered rows somewhere in your code (e.g., a slice like x_data = x_data[:100] that cuts off the last 9 rows)
• A CSV that has fewer rows than you think (open dump2.csv manually to confirm it has 109 data rows)

2. Check for Missing Values (NaN/None) #

3D plotting libraries often silently skip rows with missing numerical values. To verify this, use NumPy to check for NaNs in your data:

import numpy as np
print(f"Missing values in X: {np.isnan(x_data).any()}")
print(f"Missing values in Y: {np.isnan(y_data).any()}")
print(f"Missing values in Z: {np.isnan(z_data).any()}")

If you do have missing values, you can either:

• Filter out the bad rows:

  # Create a mask to keep only rows with no NaNs
  mask = ~np.isnan(x_data) & ~np.isnan(y_data) & ~np.isnan(z_data)
  x_clean = x_data[mask]
  y_clean = y_data[mask]
  z_clean = z_data[mask]
  print(f"Rows after filtering: {len(x_clean)}")  # Should be 109 if no missing values
  

• Fill missing values (if appropriate for your data):

  x_data = np.nan_to_num(x_data, nan=0)  # Replace NaNs with 0 (adjust value as needed)
  y_data = np.nan_to_num(y_data, nan=0)
  z_data = np.nan_to_num(z_data, nan=0)
  

3. Fix Axes Limits That Cut Off Data #

Sometimes all your data is plotted, but the default axes range is too narrow, hiding some points. Force the plot to show the full range of your data:

import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D

fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')

# Plot your data first
ax.scatter(x_clean, y_clean, z_clean)

# Set axes to cover the full min/max of your data
ax.set_xlim(np.min(x_clean), np.max(x_clean))
ax.set_ylim(np.min(y_clean), np.max(y_clean))
ax.set_zlim(np.min(z_clean), np.max(z_clean))

# Or let matplotlib auto-scale (sometimes more reliable)
ax.auto_scale_xyz(x_clean, y_clean, z_clean)

plt.show()

4. Avoid Looping to Plot Individual Points #

If you’re using a loop to plot each point one by one (e.g., for i in range(100): ax.scatter(x[i], y[i], z[i])), you might have a loop that stops early. Instead, use vectorized plotting to draw all points at once—it’s faster and less error-prone:

# Good: Plot all points in one call
ax.scatter(x_clean, y_clean, z_clean, c='red', marker='o')

# Bad: Risk of incomplete plotting
# for i in range(len(x_clean)):
#     ax.scatter(x_clean[i], y_clean[i], z_clean[i])

Full Working Example #

Here’s a consolidated code snippet that incorporates all these checks:

import numpy as np
import csv
from mpl_toolkits.mplot3d import Axes3D
import matplotlib.pyplot as plt

# Load data from dump2.csv
x_data = []
y_data = []
z_data = []

with open('dump2.csv', 'r') as f:
    reader = csv.reader(f)
    header = next(reader)  # Skip header (remove if your CSV has no header)
    for row in reader:
        # Adjust indices to match your CSV's column positions for x/y/z
        try:
            x = float(row[0])
            y = float(row[1])
            z = float(row[2])
            x_data.append(x)
            y_data.append(y)
            z_data.append(z)
        except ValueError:
            print(f"Skipping invalid row: {row}")

# Convert to NumPy arrays
x_data = np.array(x_data)
y_data = np.array(y_data)
z_data = np.array(z_data)

# Verify row count
print(f"Total rows loaded: {len(x_data)}")

# Clean missing values
mask = ~np.isnan(x_data) & ~np.isnan(y_data) & ~np.isnan(z_data)
x_clean, y_clean, z_clean = x_data[mask], y_data[mask], z_data[mask]

# Create 3D plot
fig = plt.figure(figsize=(10, 7))
ax = fig.add_subplot(111, projection='3d')

ax.scatter(x_clean, y_clean, z_clean, c='blue', alpha=0.7, marker='^')

# Auto-scale axes to show all data
ax.auto_scale_xyz(x_clean, y_clean, z_clean)

# Add labels for clarity
ax.set_xlabel(header[0] if 'header' in locals() else 'X Axis')
ax.set_ylabel(header[1] if 'header' in locals() else 'Y Axis')
ax.set_zlabel(header[2] if 'header' in locals() else 'Z Axis')

plt.title("Full 3D Plot of All 109 Cases")
plt.show()

Try these steps one by one—chances are one of them will get all your 109 cases showing up in the plot!

内容的提问来源于stack exchange，提问作者John Wilkins