TypeError: unsupported operand type(s) for -: 'NoneType' and 'float'

这个错误通常是由于在计算过程中出现了None值，无法与float类型进行计算导致的。您可以在计算之前先判断一下值是否为None，如果是None则跳过计算。以下是修改后的代码示例：

import pandas as pd
import pymysql

# 连接数据库
conn = pymysql.connect(host='localhost', port=3306, user='root', password='123456', db='daily')

# 获取所有表格名称
cursor = conn.cursor()
cursor.execute("show tables;")
tables = [table[0] for table in cursor.fetchall()]

# 遍历所有表格
for table in tables:
    # 查询net_mf_amount列的数据，按照trade_date降序排列
    sql = "select trade_date, net_mf_amount from `{}` order by trade_date desc;".format(table)
    df = pd.read_sql(sql, conn)

    # 对查询结果进行计算，将结果保存在a和b列中
    if 'a' not in df.columns:
        df['a'] = pd.Series([None]*len(df))
        df['b'] = pd.Series([None]*len(df))
    for i in range(len(df)):
        if i == 0:
            df.at[i, 'a'] = None
            df.at[i, 'b'] = None
        else:
            if pd.isna(df.at[i, 'a']) and pd.isna(df.at[i, 'b']):
                if pd.isna(df.at[i-1, 'net_mf_amount']) or pd.isna(df.at[i, 'net_mf_amount']):
                    continue
                df.at[i, 'a'] = df.at[i-1, 'net_mf_amount'] - df.at[i, 'net_mf_amount']
                if pd.isna(df.at[i-1, 'a']):
                    continue
                df.at[i, 'b'] = df.at[i-1, 'a'] - df.at[i, 'a']

    # 保存a和b列的值
    cursor.execute("alter table `{}` add column if not exists a varchar(50);".format(table))
    cursor.execute("alter table `{}` add column if not exists b varchar(50);".format(table))
    for i in range(len(df)):
        if pd.isna(df.at[i, 'a']) or pd.isna(df.at[i, 'b']):
            continue
        sql = "update `{}` set a = '{}', b = '{}' where trade_date = '{}';".format(table, df.at[i, 'a'], df.at[i, 'b'], df.at[i, 'trade_date'])
        cursor.execute(sql)

# 提交修改并关闭连接
conn.commit()
cursor.close()
conn.close()

在代码中，我们加入了判断语句，当出现None值时，直接跳过计算即可。