Python3 Nosql数据库操作包PyMongo

http://api.mongodb.com/python/current/

集合是一组文档，MongoDB中的文档类似于关系型数据库中的行，集合如同关系型数据库中的表。

集合是无模式的，一个集合里面的文档可以是各式各样的。一个集合中不光可以是值的类型不同，键也可以完全不一样，集合里可以放置任何文档。

连接数据库

import pickle
import dataset
import datetime
import numpy as np
import pandas as pd
from multiprocessing.dummy import Pool as ThreadPool
from pymongo import MongoClient
import pymongo
common_settings = {
    "serverSelectionTimeoutMS": 1000 * 1000,  
    "connectTimeoutMS": 10 * 1000,  
    "socketTimeoutMS": 1000 * 1000,  
    "maxPoolSize": 20,
    "waitQueueMultiple": 10,
}

client = MongoClient('dds-2ze6b4728dd8a1541.mongodb.rds.aliyuncs.com:3717',
                     readPreference='secondaryPreferred', **common_settings)   #replicaSet='mgset-4316217', 
# replicaSet+readPreference 使得副本集生效的地方

db = client.get_database('Database')   #mongo数据库名
db.authenticate('User Name', 'Password')   #用户名+密码
print(db.name)

collection=db['collection name']   #数据库里某个文档名

拉取指定数据库集合内的数据

import json
def get_info_from_mongo(customer_id,coll,one_fun):   
    #consume_id是DataFrame数据表,coll是mongo库的集合,one_fun是对mongo取回数据处理的函数
    re={}
    for idx in customer_id.index:
        
        cid=customer_id.iloc[idx]['cid']   #遍历用户注册id
        cid=int(cid)
        #print(cid)
        op_condition = {'customerId': {'$eq':cid}}   #条件eq等于
        #注意参数cid数据格式【int/str】与mongo库里匹配参数对应
        ret_filter = {'familyContacts':1,'companyName':1}   #与mongo库字典对应的字典格式
        res1=list(coll.find(op_condition, ret_filter))   ###find====>Query the database.
        
        resf=[]   #对idx用户的所有订单进行提取，存放到resf列表里
        for i in range(len(res1)):
            resf.append(one_fun(res1[i]))
            
        re[cid]=resf     
    return re

#test:
get_info_from_mongo(cid[:1],collection,one_fun)

向指定数据库集合内插入数据

for i in range(len(pairs)):
    cida=pairs.iloc[i,0]
    cidb=pairs.iloc[i,1]
    print(cida,cidb)
    
    b_name_match_a=get_six_character(pairs,i)['kg_b_name_match_a']
    b_phone_match_a=get_six_character(pairs,i)['kg_b_phone_match_a']
    b_relationship_match_a=get_six_character(pairs,i)['kg_b_relationship_match_a']
    
    a_name_match_b=get_six_character(pairs,i)['kg_a_name_match_b']
    a_phone_match_b=get_six_character(pairs,i)['kg_a_phone_match_b']
    a_relationship_match_b=get_six_character(pairs,i)['kg_a_relationship_match_b']

    coll.update_one({"$and":[{'a_customer_id':int(cida)},{'b_customer_id':int(cidb)}]},
                    {'$set': {"relation_feature.kg_b_name_match_a":b_name_match_a,
                              "relation_feature.kg_b_phone_match_a":b_phone_match_a,
                              "relation_feature.kg_b_relationship_match_a":b_relationship_match_a,
                              "relation_feature.kg_a_name_match_b":a_name_match_b,
                              "relation_feature.kg_a_phone_match_b":a_phone_match_b,
                              "relation_feature.kg_a_relationship_match_b":a_relationship_match_b}}, 
                    upsert=True)

{
    "_id" : ObjectId("5b87c3c62bb890b992659017"),
    "a_customer_id" : 236,
    "b_customer_id" : 2560842,
    "relation_feature" : {
        "kg_address_match_cnt" : 0,   ###使用update_one插入，不覆盖原本有的别的同一层数据
        "kg_b_name_match_a" : 0,
        "kg_b_phone_match_a" : 0,
        "kg_b_relationship_match_a" : [],
        "kg_a_name_match_b" : 0,
        "kg_a_phone_match_b" : 0,
        "kg_a_relationship_match_b" : [],
    }
}