一直以为csv靠逗号（,）分割数据，那么只要用str.spilt(',',row)便可以将数据分割开。

事实证明想简单了，csv里还有这样的规定，如果数据内有双引号（"）和逗号（,）那么，数据就要用双引号括起来，而双引号要替换两个双引号。

比如说数据是: hello,world，那么他在csv里应该表达为："hello,world"

数据是: say "hello"，那么他在csv里就应该是"say ""hello"""

这样就使得解析的时候没法直接用split函数。

根据规则解析应该符合以下状态图：

（有些字倒，将就看吧）

那么使用python实现便是：

class CsvInterpreter:_lines = {"0": {"quot": ("1", False), "other": ("2", True)},"1": {"other": ("1", True), "quot": ("3", False), "comma": ("1", True)},"2": {"other": ("2", True), "comma": ("0", False), "enter": ("4", False)},"3": {"quot": ("1", True), "comma": ("0", False), "enter": ("4", False)},"5": {"comma": ("0", False), "enter": ("4", False)}}def _init(self):self._chars = ""self._current_status = "0"self._buffer = ""self._container = []def __init__(self):self._buffer = Noneself._chars = Noneself._container = Noneself._current_status = Noneself._init()def _next(self, char: str) -> None:if char == '"':cond = "quot"elif char == ',':cond = "comma"elif char == '\n':cond = "enter"else:cond = "other"if cond not in self._lines[self._current_status]:raise ValueError("格式不正确", self._chars, self._container, self._buffer, self._current_status, cond, char)else:next_status, if_input = self._lines[self._current_status][cond]self._current_status = next_statusif if_input:self._buffer += chardef split(self, line: str) -> list:self._init()self._chars = linefor char in line:self._next(char)if self._current_status in ["0", "4"]:self._container.append(self._buffer)self._buffer = ""if self._current_status == "4":return self._containerif self._current_status not in ["3", "4"]:raise ValueError("格式不正确")self._container.append(self._buffer)return self._container

运行：

csv_interpreter = CsvInterpreter()
with open('*.csv','r',encoding='utf-8') as f:for line in f.readlines():row = csv_interpreter.split(line)print(row)