如何遍历气象数据目录文件，计算各站点月总降水量并输出CSV

Got it, let's build a Python script that handles this precipitation data aggregation task perfectly. I'll break it down step by step so you can follow along and adjust it to your exact CSV structure:

First, we'll use pandas for this task—it's ideal for handling CSV data and aggregating totals. If you don't have it installed yet, run this in your terminal:

pip install pandas

The Script #

import pandas as pd
from pathlib import Path

# Point to your database directory on the Desktop
data_directory = Path.home() / "Desktop" / "database"

# Empty list to collect all processed daily records
daily_records = []

# Loop through every CSV file in the directory
for csv_file in data_directory.glob("*.csv"):
    print(f"Processing: {csv_file.name}")
    
    # Read the CSV, extracting only the columns we need
    # --- IMPORTANT ADJUSTMENTS HERE ---
    # Assumptions about your CSV structure (0-based indices):
    # - Index 0: Station ID/Name (e.g., "Station_A", "Station_1")
    # - Index 1: Date (format like YYYY-MM-DD or YYYY/MM/DD)
    # - Index 16: Precipitation value (your target column)
    daily_data = pd.read_csv(
        csv_file,
        usecols=[0, 1, 16],
        header=None,  # Delete this line if your CSV has a header row
        names=["Station", "Date", "Precipitation"]  # Name the columns for clarity
    )
    
    # Convert date column to datetime to extract month
    daily_data["Date"] = pd.to_datetime(daily_data["Date"], errors="coerce")
    
    # Remove rows with invalid dates (if any)
    daily_data = daily_data.dropna(subset=["Date"])
    
    # Extract month number (1 = January, 12 = December)
    daily_data["Month"] = daily_data["Date"].dt.month
    
    # Add this file's data to our master list
    daily_records.append(daily_data)

# Combine all daily data into one DataFrame
all_data = pd.concat(daily_records, ignore_index=True)

# Calculate total precipitation per station per month
monthly_totals = all_data.groupby(["Station", "Month"], as_index=False)["Precipitation"].sum()

# Rename the sum column for readability
monthly_totals.rename(columns={"Precipitation": "Total_Precipitation"}, inplace=True)

# Sort results by station and month to get clean 48-row output (4 stations × 12 months)
monthly_totals = monthly_totals.sort_values(by=["Station", "Month"]).reset_index(drop=True)

# Save the final result to a CSV on your Desktop
output_file = Path.home() / "Desktop" / "monthly_precipitation_summary.csv"
monthly_totals.to_csv(output_file, index=False)

print(f"Done! Your 48-row summary is saved to: {output_file}")

Key Adjustments for Your Data #

• CSV Header Row: If your CSV files have a header row (e.g., first line says "Station,Date,Temperature,..."), delete the header=None line and update usecols to use column names instead of indices (e.g., usecols=["Station", "ObservationDate", "DailyPrecip"]).
• Non-Standard Date Format: If your dates are in a weird format (like DD/MM/YYYY), add the format parameter to pd.to_datetime: pd.to_datetime(daily_data["Date"], format="%d/%m/%Y", errors="coerce").
• Station in Filename: If your CSV files don't have a station column but the filename includes the station ID (e.g., "station_2_20230515.csv"), extract it from the filename like this:

  # Add this right after opening the CSV file
  station_id = csv_file.name.split("_")[1]  # Adjust split logic to match your filename pattern
  daily_data["Station"] = station_id
  

This script will process all your CSV files, aggregate the monthly totals for each of the 4 stations, and output a clean 48-row CSV with the results.

内容的提问来源于stack exchange，提问作者Hun_