python_0">一、python的数据表处理

python">"""
删除 Doris库中某些表中无效的数据
"""
import mysql.connector
import socket
import socks
import pandas as pd
import pymysql
from sqlalchemy import create_engine, text
import csv
from datetime import datetime# Tidb数据库
def DorisTableInfo(databaseName, env):if env == "online":# Dorisdb = pymysql.connect(host="yourHost",port=6666,user="yourUser",password="yourPwd",database=databaseName)cursor = db.cursor()cursor.execute("set names utf8")cursor.execute("SET character_set_connection=utf8;")dbName = "MY_DB.tableName"sql="select id from %s where status = 'InValid' order by id asc limit 500000;" % (dbName)print("查询要删除的数据语句：", sql)cursor.execute(sql)data = cursor.fetchall()print("查询 %s 需要物理删除的数据量: " % (dbName), len(data))list = []for i in range(0, len(data)):list.append(data[i][0])# print(list)# 1000个一拆分listItem=group_list(list, 1000)for i in range(len(listItem)):deleteIds = ','.join(str(item) for item in listItem[i])deleteSql = "delete from %s where  id in (%s)" % (dbName, deleteIds)print("[%s/%s]" % (i, len(listItem)), deleteSql)cursor.execute(deleteSql)def group_list(lst, group_size):groups = {}for i, item in enumerate(lst):group_number = i // group_sizeif group_number not in groups:groups[group_number] = []groups[group_number].append(item)return [group for _, group in sorted(groups.items())]# Doris 数据库
databaseName = "MY_DB"env = "online"
print("start-------")
DorisTableInfo(databaseName, env)
print("done-------")