Python处理结构化数据的12个核心模块全解析

import csv
# 写入带标题的表格数据
data = [
    ["电影名称", "票房(万)", "排片占比"],
    ["流浪地球3", "45200", "32.4%"],
    ["唐人街探案4", "38700", "28.1%"]
]
with open("box_office.csv", "w", encoding="utf-8", newline="") as f:
    writer = csv.writer(f, dialect="excel")  # 使用Excel风格
    writer.writerows(data)

import pandas as pd
# 从CSV创建DataFrame并处理缺失值
df = pd.read_csv("box_office.csv", parse_dates=["上映日期"])
df["票房(亿)"] = df["票房(万)"].astype(float) / 10000
df.fillna(0, inplace=True) # 填充缺失值
df.to_excel("processed_data.xlsx", index=False) # 输出Excel

total = df["票房(万)"].sum()
df["市场占比"] = (df["票房(万)"] / total * 100).round(2)

import pickle
# 序列化包含自定义类的对象
class Movie:
    def __init__(self, title, box_office):
        self.title = title
        self.box_office = box_office

movie = Movie("哪吒2", 568000)
with open("movie.pkl", "wb") as f:
    pickle.dump(movie, f, protocol=pickle.HIGHEST_PROTOCOL)  # 使用最高协议版本

import shelve
# 存储电影评分数据
with shelve.open("movie_ratings.db") as db:
    db["哪吒2"] = {"score": 9.5, "votes": 125000}
    db["唐探4"] = {"score": 8.7, "votes": 98000}
    # 自动处理数据持久化

import struct
# 解析BMP文件头（14字节）
with open("image.bmp", "rb") as f:
    header = f.read(14)
    # '2sIHH'表示：2字节字符串+无符号整数+2个无符号短整数
    magic, file_size, reserved1, reserved2 = struct.unpack("<2sIHH", header)
    if magic != b"BM":
        raise ValueError("非BMP文件")

import configparser
# 写入多层级配置
config = configparser.ConfigParser()
config["DEFAULT"] = {
    "retry_times": "3",
    "timeout": "10"
}
config["DATABASE"] = {
    "host": "127.0.0.1",
    "port": "5432"
}
with open("settings.ini", "w") as f:
    config.write(f)

import xml.etree.ElementTree as ET
# 生成符合规范的XML
root = ET.Element("movies")
movie = ET.SubElement(root, "movie", attrib={"id": "1001"})
ET.SubElement(movie, "title").text = "封神第二部"
ET.SubElement(movie, "year").text = "2025"
tree = ET.ElementTree(root)
tree.write("movies.xml", encoding="utf-8", xml_declaration=True)

import numpy as np
# 创建评分矩阵（用户×电影）
ratings = np.array([
    [9.2, 8.5, np.nan],
    [8.7, np.nan, 9.0],
    [np.nan, 7.8, 8.9]
])
# 计算每部电影的平均分（忽略NaN）
mean_ratings = np.nanmean(ratings, axis=0)
print(f"电影平均分：{mean_ratings.round(1)}")

import array
# 存储百万级票房数据
box_offices = array.array("f", [45.2, 38.7, 56.8])  # 'f'表示单精度浮点数
box_offices.extend([32.1, 47.9])
print(f"总票房：{sum(box_offices):.1f}亿")

import sqlite3
# 创建内存数据库分析票房数据
conn = sqlite3.connect(":memory:")
cursor = conn.cursor()
cursor.execute("""
    CREATE TABLE movies (
        id INTEGER PRIMARY KEY,
        title TEXT,
        box_office REAL,
        release_date DATE
    )
""")
# 批量插入数据
movies = [
    ("哪吒2", 56.8, "2025-02-12"),
    ("唐探4", 38.7, "2025-01-29")
]
cursor.executemany("INSERT INTO movies VALUES (NULL, ?, ?, ?)", movies)
conn.commit()
# 查询票房前3的电影
for row in cursor.execute("SELECT * FROM movies ORDER BY box_office DESC LIMIT 3"):
    print(row)
conn.close()

import msgpack
# 序列化电影数据（需安装msgpack-python）
movie_data = {
    "title": "流浪地球3",
    "daily_box_office": [4520, 3870, 5680],
    "is_showing": True
}
packed = msgpack.packb(movie_data, use_bin_type=True)
unpacked = msgpack.unpackb(packed)
print(f"反序列化结果：{unpacked['title']}")

chunk_size = 10000
for chunk in pd.read_csv("large_data.csv", chunksize=chunk_size):
    process(chunk)  # 处理每个数据块