需求：领导要求分析服务日志，计算出各个设备的在线时长，在线率，设备使用率等。
服务日志格式为：
2018-03-01 00:13:52,815 [protocol.handler.1][INFO] - cn.testin.trans.controller.req.script.HeartBeat.deviceLog---{"sdcardWriteable":0,"deviceAction":1,"sdcardAvail":10113417216,"wifiSsid":"testin03_default","rootEnable":0,"deviceid":"863563023966636","networkState":1,"errorMsg":"","minicapEnable":1,"deviceState":1,"romAvail":1655619584,"rpiid":"localhost","networkType":1,"ramAvail":1488954688,"debugMode":1,"errorCount":0,"batteryLevel":100}
以上标红部分：
deviceid: 为设备id， networkState:为网络状态，该值有1、0两个值 deviceState:为设备状态，该值有0、1、2三个值
device_report.py
#!/usr/bin/env python
# -*- coding: utf-8 -*-
import sys
import json
from collections import Counter
from mysql import connector
import datetime
today = datetime.datetime.now().strftime("%Y%m%d")
def execute_sql(insert_sql) :
try:
# 配置信息
config = {
'host': '127.0.0.1',
'port': 3306,
'user': 'analy',
'password': '****',
'database': 'analy',
'charset': 'utf8'
}
# 连接
con = connector.connect(**config)
cursor = con.cursor()
# 利用字符串方式查询
cursor.execute(insert_sql)
con.commit()
# 关闭
cursor.close()
con.close()
except connector.Error, e:
print(e)
print(e.message)
#cur = conn.cursor()
records = None
def device():
with open(sys.argv[1]) as f:
for line in f:
#取出日志里的json字符串
json_origin = line.split('---')[1]
#将json字符串格式化为字典
json_final = json.loads(json_origin)
DeviceId = json_final["deviceid"]
NetworkState = json_final["networkState"]
DeviceState = json_final["deviceState"]
yield [DeviceId,NetworkState,DeviceState]
DeviceTotal,NetworkState_0,NetworkState_1,DeviceState_0,DeviceState_1,DeviceState_2 = Counter(), Counter(), Counter(), Counter(), Counter(), Counter()
header = ['deviceid','NetworkState','DeviceState']
for line in device():
datadict = dict(zip(header,line))
DeviceTotal[datadict['deviceid']] +=1
if datadict['NetworkState'] == 0:
NetworkState_0[datadict['deviceid']] +=1
elif datadict['NetworkState'] == 1:
NetworkState_1[datadict['deviceid']] +=1
if datadict['DeviceState'] == 0:
DeviceState_0[datadict['deviceid']] +=1
elif datadict['DeviceState'] == 1:
DeviceState_1[datadict['deviceid']] +=1
elif datadict['DeviceState'] == 2:
DeviceState_2[datadict['deviceid']] +=1
if DeviceTotal:
for k,v in DeviceTotal.most_common(records):
data = (today,k, (NetworkState_0[k]/float(v))*100, (NetworkState_1[k]/float(v))*100, (DeviceState_0[k]/float(v))*100, (DeviceState_1[k]/float(v))*100, (DeviceState_2[k]/float(v))*100,((DeviceState_1[k] + DeviceState_0[k])*10))
query = ("INSERT INTO device_data_report""(date,deviceid,netstate_0_rate,netstate_1_rate,devicestate_0_rate,devicestate_1_rate,devicestate_2_rate,Online_time)" "VALUES('%s','%s','%s','%s','%s','%s','%s','%s')") % data
execute_sql(query)此时执行完成，需要不到两分钟。
此前shell脚本：
#!/bin/bash
#show all deviceid counts
Today=`date "+%Y%m%d"`
#FileName='test.txt'
FileName='devicereport'
TempFile='count1.txt'
mysql_sql='mysql -uanaly -p****** '
cat ${FileName} | awk -F'---' '{print $2}' | jq '.deviceid'|sort|uniq -c >> ${TempFile}
for deviceid in `cat ${TempFile}|awk '{print $2}'`
do
#find networkstate
netstate_0=`grep $deviceid ${FileName}| awk -F'---' '{print $2}' | jq '.networkState'|grep 0 |wc -l`
netstate_1=`grep $deviceid ${FileName}| awk -F'---' '{print $2}' | jq '.networkState'|grep 1 |wc -l`
devicestate_0=`grep $deviceid ${FileName}| awk -F'---' '{print $2}' | jq '.deviceState'|grep 0 |wc -l`
devicestate_1=`grep $deviceid ${FileName}| awk -F'---' '{print $2}' | jq '.deviceState'|grep 1 |wc -l`
devicestate_2=`grep $deviceid ${FileName}| awk -F'---' '{print $2}' | jq '.deviceState'|grep 2 |wc -l`
sumdevice=`grep $deviceid $TempFile|awk '{print $1}'`
#compute rate
netstate_0_rate=`echo "scale=2;a=(${netstate_0}/${sumdevice})*100;if(length(a)==scale(a)) print 0;print a"|bc`
netstate_1_rate=`echo "scale=2;a=(${netstate_1}/${sumdevice})*100;if(length(a)==scale(a)) print 0;print a"|bc`
devicestate_0_rate=`echo "scale=2;a=(${devicestate_0}/${sumdevice})*100;if(length(a)==scale(a)) print 0;print a"|bc`
devicestate_1_rate=`echo "scale=2;a=(${devicestate_1}/${sumdevice})*100;if(length(a)==scale(a)) print 0;print a"|bc`
devicestate_2_rate=`echo "scale=2;a=(${devicestate_2}/${sumdevice})*100;if(length(a)==scale(a)) print 0;print a"|bc`
Online_time=$(((${devicestate_0} + ${devicestate_1})/360))
${mysql_sql} -e 'insert into analy.device_data_report (date,deviceid,netstate_0_rate,netstate_1_rate,devicestate_0_rate,devicestate_1_rate,devicestate_2_rate,Online_time) values ('$Today','$deviceid','$netstate_0_rate','$netstate_1_rate','$devicestate_0_rate','$devicestate_1_rate','$devicestate_2_rate','$Online_time')'此时执行完需要两个多小时。
脚本执行完入库的情况如下图：