如何用Python高效生成命名规范的多DataFrame用于论文数据分析

Hey there! Great question—when dealing with multiple DataFrames tied to participants and their tests, creating individual variables like df11, df12 can get messy quickly, especially if you need to access or manipulate them in bulk. Let’s break down the best approaches to handle this efficiently.

Recommended: Use a Dictionary to Store DataFrames #

This is the cleanest, most scalable method. A dictionary lets you map your dfXX-style labels to their corresponding DataFrames, making it easy to access individual frames or batch-process all of them.

Here’s how to implement it:

import pandas as pd

# Define your participant and test counts (adjust these to your actual numbers)
total_participants = 15
total_tests = 3

# Initialize an empty dictionary to hold all DataFrames
df_collection = {}

for p in range(total_participants):
    participant_num = p + 1
    for t in range(total_tests):
        test_num = t + 1
        # Generate your filename (using f-strings for readability)
        filename = f"P{participant_num}S{test_num}.csv"
        # Create the dfXX-style key (e.g., df11, df12)
        df_key = f"df{participant_num}{test_num}"
        # Read the CSV and store it in the dictionary
        df_collection[df_key] = pd.read_csv(filename)

How to Access & Use the DataFrames #

• To pull a specific DataFrame (e.g., participant 1, test 1):

  specific_df = df_collection["df11"]
  

• To batch-process all DataFrames (e.g., remove missing values from every frame):

  for df_label, df in df_collection.items():
      # Modify the DataFrame in-place or update the dictionary entry
      df.dropna(inplace=True)
      # Or if you prefer not to modify in-place:
      # df_collection[df_label] = df.dropna()
  

Why this works:

• No cluttered namespace: You won’t have dozens of dfXX variables floating around in your environment.
• Easy bulk operations: Loop through the dictionary to apply transformations, run analyses, or export results across all DataFrames.
• Clear organization: Keys make it obvious which DataFrame corresponds to which participant/test pair.

Alternative: Create Individual Global Variables (Not Recommended) #

If you absolutely need standalone variables like df11 (e.g., for quick one-off checks), you can use Python’s globals() function to inject them into the global namespace. However, this is not ideal for long-term maintainability.

import pandas as pd

total_participants = 15
total_tests = 3

for p in range(total_participants):
    participant_num = p + 1
    for t in range(total_tests):
        test_num = t + 1
        filename = f"P{participant_num}S{test_num}.csv"
        df_key = f"df{participant_num}{test_num}"
        # Create a global variable with the dfXX name
        globals()[df_key] = pd.read_csv(filename)

Caveats of This Approach #

• You can’t easily iterate over all dfXX variables for bulk tasks (you’d have to manually track all names).
• Risk of overwriting existing variables if your dfXX labels clash with other names in your code.
• Makes debugging and code maintenance harder as your project grows.

Bonus: Nested Dictionary for Participant-First Organization #

If you want to group DataFrames by participant first (instead of flat dfXX keys), a nested dictionary adds extra structure:

participant_groups = {}

for p in range(total_participants):
    participant_num = p + 1
    # Create a sub-dictionary for each participant
    participant_groups[participant_num] = {}
    for t in range(total_tests):
        test_num = t + 1
        filename = f"P{participant_num}S{test_num}.csv"
        participant_groups[participant_num][test_num] = pd.read_csv(filename)

# Access participant 3's test 2 DataFrame
participant_3_test_2 = participant_groups[3][2]

This is especially useful if you plan to analyze data per participant before aggregating across tests.

内容的提问来源于stack exchange，提问作者Adon Alves