CSV 文件

以文本形式保存的表格

CSV（逗号分隔值）文件把表格存成文本：一行一条记录，单元格用逗号分隔，通常有一行表头给出列名。价格历史、成交明细以及几乎所有工具的导出都是这种格式。看起来用 line.split(",") 就能解析，但单元格本身可能含有逗号，这时它会被引号包起来：AAPL,"Q3, adjusted"。手工切分在这里会出错，所以请使用标准库的 csv 模块，并用 newline="" 打开文件，以便正确处理单元格内的换行。

import csv

# Write a small table with csv.writer
with open("prices.csv", "w", newline="", encoding="utf-8") as file:
    writer = csv.writer(file)
    writer.writerow(["date", "symbol", "close", "note"])
    writer.writerow(["2025-01-02", "AAPL", 184.25, "Q3, adjusted"])
    writer.writerow(["2025-01-03", "AAPL", 185.5, ""])

# The comma inside the note is quoted for us
with open("prices.csv", encoding="utf-8") as file:
    print(file.read().splitlines()[1])  # 2025-01-02,AAPL,184.25,"Q3, adjusted"

读取行

csv.reader(file) 把每一行作为字符串列表返回。csv.DictReader(file) 先读表头，再把每一行作为以列名为键的字典返回，可读性更好，列顺序变化也不受影响。无论哪种方式，每个单元格都是文本：计算前要用 float(...) 或 int(...) 转换。对使用其他分隔符的文件，传入 delimiter=";"（或 "\t"）。

import csv

# DictReader gives each row as a dict; cells are strings
closes = []
with open("prices.csv", newline="", encoding="utf-8") as file:
    for row in csv.DictReader(file):
        closes.append(float(row["close"]))
print(closes)  # [184.25, 185.5]

# csv.reader gives plain lists, header included
with open("prices.csv", newline="", encoding="utf-8") as file:
    rows = list(csv.reader(file))
print(rows[0])  # ['date', 'symbol', 'close', 'note']
print(rows[1])  # ['2025-01-02', 'AAPL', '184.25', 'Q3, adjusted']

写入行

csv.writer(file).writerow(values) 写入一行，并在需要时加引号，数字会自动转成文本。csv.DictWriter(file, fieldnames=[...]) 写入字典，writeheader() 先写列名。用 "w" 和 newline="" 打开文件；没有 newline="" 时，Windows 上行与行之间可能出现空行。

import csv

# DictWriter: header first, then one dict per row
trades = [{"symbol": "AAPL", "quantity": 200}, {"symbol": "MSFT", "quantity": 50}]
with open("trades.csv", "w", newline="", encoding="utf-8") as file:
    writer = csv.DictWriter(file, fieldnames=["symbol", "quantity"])
    writer.writeheader()
    writer.writerows(trades)

with open("trades.csv", encoding="utf-8") as file:
    print(file.read().splitlines())  # ['symbol,quantity', 'AAPL,200', 'MSFT,50']

import csv

# Cells are text until you convert them
with open("prices.csv", newline="", encoding="utf-8") as file:
    row = next(csv.DictReader(file))
print(repr(row["close"]), row["close"] * 2)   # '184.25' 184.25184.25
print(float(row["close"]) * 2)                # 368.5

# Another separator: semicolons
with open("euro.csv", "w", newline="", encoding="utf-8") as file:
    csv.writer(file, delimiter=";").writerow(["AAPL", "184,25"])
with open("euro.csv", newline="", encoding="utf-8") as file:
    print(next(csv.reader(file, delimiter=";")))  # ['AAPL', '184,25']

• 用 newline="" 打开；不要手工拆分或拼接 CSV 行。
• csv.DictReader 把行读成字典；每个单元格都是字符串。
• csv.writer / csv.DictWriter 写入行并自动加引号。

下一节：JSON 数据，用于无法放进平面表格的嵌套记录。

实验：评分内容

• 用 read_closes 按代码归组收盘价25 分
• 用 write_trades 把成交保存为 CSV 文件25 分
• 用 total_by_symbol 汇总每个代码的成交金额25 分
• 用 filter_rows 把某个代码的行复制到新文件25 分