import base64
import time
import traceback
import urllib.request
import gevent
from gevent import monkey
import sys
monkey.patch_all()
import schedule
from gevent import pool
from pymongo import MongoClient
from datetime import datetime, timedelta
import logging
import json
import requests
import os
import uuid
from PIL import Image
import io
from loguru import logger
client = MongoClient('mongodb://username:password@127.0.0.1:27017/')
# 库 - database
client_db = client['data_base']
# 表 - collection
temp_data_col = client_db['temp_data']
def work_func():
# 检索条件
mongo_condition_dict = {
"create_time": {"$gte": "2021-04-15T00:37:46.110000"}
}
# 过滤返回的字段
filter_colum_dict = {
"_id": 0, "username": 1, "password": 1
}
detail_list =temp_data_col.find(mongo_condition_dict, filter_colum_dict).batch_size(2000)
logger.info('Start search Count Is {}'.format(detail_list.count()))
temp_data_set = set()
for detail_info indetail_list:
logger.info(f"detail_info = {detail_info}")
username, password = detail_info["username"], detail_info["password"]
temp_data_set.add(username + '----' + password)
logger.info(f"查询到数据: = {temp_data_set}")
logger.info(f"查询到总数: = {len(temp_data_set)}")
# 按行写入文件,如何优化?一次打开,按行写入 - 待优化
for data in temp_data_set:
with open('./temp.txt', 'a') as f:
f.write(data + '\n')
if __name__ == '__main__':
work_func()
复制代码
© 版权声明
文章版权归作者所有,未经允许请勿转载。
THE END