如何优雅优化Pandas循环中带编号命名的统计新列创建代码

Nice job getting the original script working! Let’s clean this up to make it more maintainable, concise, and less prone to typos (like the kurt_col mistakenly using mean_ in your original code). Here’s a streamlined approach:

Step 1: Centralize Your Slice Ranges #

Instead of manually creating each phen_* DataFrame, store your column ranges in a list. This makes it easy to adjust ranges or add new segments later:

# Define all column slice ranges in one place
slice_ranges = [(0, 20), (20, 55), (55, 70), (70, 80), (70, 110)]

Step 2: Use a Dictionary for Statistical Metrics #

Map the names you want for your new columns to the corresponding Pandas methods. This eliminates repetitive code and makes adding new metrics trivial:

# Map desired column prefixes to Pandas row-wise methods
stats_mapping = {
    "mean": "mean",
    "std": "std",
    "skew": "skew",
    "kurt": "kurt",
    "total": "sum"
}

Step 3: Streamline the Loop with enumerate #

Use enumerate to automatically track your phen number (starting at 1) instead of manually incrementing a variable. Combine this with the above mappings to generate all columns cleanly:

for phen_num, (start, end) in enumerate(slice_ranges, start=1):
    # Extract the current segment of rain_calc
    sub_df = rain_calc.iloc[:, start:end]
    # Calculate each metric and add to original_df
    for stat_name, method in stats_mapping.items():
        col_name = f"{stat_name}_{phen_num}"
        original_df[col_name] = getattr(sub_df, method)(axis=1)

Why This Works Better #

• Maintainability: All slice ranges and metrics are in one place—no need to hunt through code to adjust values or add new calculations.
• Error Reduction: No more manual string concatenation for column names (which caused your kurt_col typo) or manual counter management.
• Scalability: Want to add a new metric like median? Just add "median": "median" to stats_mapping. Need a new column segment? Add a tuple to slice_ranges.
• Readability: The code clearly expresses intent—anyone reading it can quickly see what segments and metrics are being processed.

内容的提问来源于stack exchange，提问作者Reut