亲们。。。我又来求助了！！

purelove_88

回复 9# yestreenstars


[2014-07-11 14:38:24] {"log_type":"ROLE_LOGIN","log_data":{"SERVER_ID":"11","ROLE_ID":"1897","CHANNEL_CODE":"uc","ROLE_LEVEL":"48","USER_ID":"CU20140710030409846","OPERATION":"12","LOGIN_DATE":"1405060670603","LOGOUT_DATE":"1405060704060","GAME_TIME":"33456","CREATE_DATE":"1404964589117"}}


两个log文件的日志格式是一样的

我想从11号的日志中找到10号同样登陆的玩家，用user_id作为标准，进行对比。

   

purelove_88

回复 10# reyleon

1：我们公司的日志格式是json的，如下
[2014-07-11 14:38:24] {"log_type":"ROLE_LOGIN","log_data":{"SERVER_ID":"11","ROLE_ID":"1897","CHANNEL_CODE":"uc","ROLE_LEVEL":"48","USER_ID":"CU20140710030409846","OPERATION":"12","LOGIN_DATE":"1405060670603","LOGOUT_DATE":"1405060704060","GAME_TIME":"33456","CREATE_DATE":"1404964589117"}}


2：我们的日志文件，是每日保存，现在有10号和11号两个日志，我想用11号的日志文件与10号的做对比，如果10号的日志中包含的userid在11号的日志中也登陆了，那么记做一次有效数据


3：我想做的就是11号和10号的日志对比，确定次日留存率

求教育啊。。。。


   

Herowinter

回复 12# purelove_88
瞎猜一个，不对再改吧。

1. awk '{uid=gensub(/.*USER_ID":"([^"]*).*/,"\\1",1)} NR==FNR{a[uid];next} {if(uid in a){b[uid];count++}} END{print "USER_ID exists in both log files:";for(i in b)print i;print "total:"count}' 1.log 2.log
   
2. USER_ID exists in both log files:
   
3. CU20140710030409846
   
4. total:1

复制代码

1. cat 1.log
   
2. [2014-07-10 14:38:24] {"log_type":"ROLE_LOGIN","log_data":{"SERVER_ID":"11","ROLE_ID":"1897","CHANNEL_CODE":"uc","ROLE_LEVEL":"48","USER_ID":"CU20140710030409846","OPERATION":"12","LOGIN_DATE":"1405060670603","LOGOUT_DATE":"1405060704060","GAME_TIME":"33456","CREATE_DATE":"1404964589117"}}
   
3. [2014-07-10 14:38:24] {"log_type":"ROLE_LOGIN","log_data":{"SERVER_ID":"11","ROLE_ID":"1897","CHANNEL_CODE":"uc","ROLE_LEVEL":"48","USER_ID":"CU20140710030409847","OPERATION":"12","LOGIN_DATE":"1405060670603","LOGOUT_DATE":"1405060704060","GAME_TIME":"33456","CREATE_DATE":"1404964589117"}}
   
4. 
   
5. cat 2.log
   
6. [2014-07-11 14:38:24] {"log_type":"ROLE_LOGIN","log_data":{"SERVER_ID":"11","ROLE_ID":"1897","CHANNEL_CODE":"uc","ROLE_LEVEL":"48","USER_ID":"CU20140710030409846","OPERATION":"12","LOGIN_DATE":"1405060670603","LOGOUT_DATE":"1405060704060","GAME_TIME":"33456","CREATE_DATE":"1404964589117"}}
   
7. [2014-07-11 14:38:24] {"log_type":"ROLE_LOGIN","log_data":{"SERVER_ID":"11","ROLE_ID":"1897","CHANNEL_CODE":"uc","ROLE_LEVEL":"48","USER_ID":"CU20140710030409848","OPERATION":"12","LOGIN_DATE":"1405060670603","LOGOUT_DATE":"1405060704060","GAME_TIME":"33456","CREATE_DATE":"1404964589117"}}
   

复制代码

reyleon

回复 12# purelove_88

1. #!/usr/bin/python
   
2. # -*- coding:utf-8 -*-
   
3. import json
   
4. import re
   
5. 
   
6. # 昨天的日志文件
   
7. ztlog = 'log0710'
   
8. allUserId = []
   
9. zt = open(ztlog).readlines()
   
10. for r in zt :
    
11.     jsonString = re.sub(r'\[.*?\]','',r)
    
12.     data = json.loads(jsonString)
    
13.     allUserId.append(data['log_data']['USER_ID'])
    
14. 
    
15. # 有效计数
    
16. sum = 0
    
17. # 今天的日志文件
    
18. jtlog = 'log0711'
    
19. jt = open(jtlog).readlines()
    
20. for r in jt :
    
21.     jsonString = re.sub(r'\[.*?\]','',r)
    
22.     data = json.loads(jsonString)
    
23.     if data['log_data']['USER_ID'] in allUserId:
    
24.         sum+=1
    
25. print sum

复制代码

Herowinter

回复 14# reyleon
处理Jason用python好点，不像咱用的是土办法，学习学习。

   

yestreenstars

回复 15# Herowinter

@jason680有人要处理你~
   

Herowinter

回复 16# yestreenstars
造得一手好谣。

   

q1208c

1. sed -e 's/^.*USER_ID":"\([^"]\+\)".*$/\1/' log1 | sort -u > log1.user
   
2. sed -e 's/^.*USER_ID":"\([^"]\+\)".*$/\1/' log2 | sort -u > log2.user
   
3. grep -c -Ff log1.user log2.user

复制代码

不知道效率如何. 数据量太少, 没法测试.

刺客阿地

回复 5# Herowinter

我也好羡慕啊。。。游戏运维大神都在30K的。。。。