Scala 经典的模式匹配和尾递归
package io
import java.io.{BufferedWriter, File, FileWriter}
import java.text.SimpleDateFormat
import scala.annotation.tailrec
import scala.io.Source
/**
* author: HuaZhe Ray
*
* describe: 计算服务调用平均时长
*
* createDate: 2017/12/15
* createTime: 10:57
*
*/
object AverageService {
type TIME = String
type THREAD = String
type SERVICE = String
type WAYS = String
type METHOD = String
val sdf = new SimpleDateFormat("MM-dd hh:mm:ss SSS")
val regex = """(\d{2}-\d{2} \d{2}:\d{2}:\d{2} \d{3}) ([^\s]+) DEBUG ([^\s]+) ([^\s]+) ([^\s]+) ([^\s]+) (.+)""".r
def main(args: Array[String]): Unit = {
// 服务调用次数
countByServiceName
Thread.sleep(1000)
// 服务调用时长
analyzeAverageTime
}
/**
* 最新优化，求服务调用次数 根据 带有request单词的作为一次服务，并过滤掉 INFO 日志， 取 DEBUG 日志
* 2017-12-14 14:50
*/
def countByServiceName(): Unit = {
val source = Source.fromFile("E:\\scala\\detail-test.log", "UTF-8")
val lists:List[(String,String,String,String,String)] = source.getLines().collect{
case regex(timestamp, thread,c,service,version,method,requests) => (timestamp,thread,service,method,requests)
}.toList
val requests:List[(String,String,String)] = lists.filter(_._5.contains("request")).map (line =>{
(line._1,line._2,line._3+"."+line._4)
})
val countByServiceName:Map[String,Int] = requests.map((_, 1)).groupBy(_._1._3).map(t=>(t._1,t._2.size))
countByServiceName.foreach {
case (service, count) => println(s"$service count:$count")
}
}
/**
* 2017-12-15 统计服务调用平均时长
*
* 思路，根据正则 截取到 request 和 response 的 服务 ，然后根据时间排序 ---> 每次调用的 response - request 得到 服务调用时长
*/
def analyzeAverageTime(): Unit = {
val source = Source.fromFile("E:\\scala\\detail-test.log", "UTF-8")
val lists:List[(TIME,THREAD,SERVICE,METHOD,WAYS)] = source.getLines().collect{
case regex(timestamp, thread,c,service,version,method,requests) => (timestamp,thread,service,method,requests)
}.toList
val requests:List[(TIME,THREAD,SERVICE,WAYS)] = lists.filter(_._5.contains("request")).map (line =>{
(line._1,line._2,line._3+"."+line._4,"request")
})
val response:List[(TIME,THREAD,SERVICE,WAYS)] = lists.filter(_._5.contains("response")).map (line =>{
(line._1,line._2,line._3+"."+line._4,"response")
})
// 得到所有 request 和 response 结果
val totalServices = (requests ++ response).sortWith((t1,t2) => t1._1 < t2._1).take(10)
//具体进行处理
processByTailRec(totalServices,Map[(THREAD,SERVICE),TIME]())
}
/**
* 通过定义 mutable Map 来 存放 foreach 临时变量， 进行相减操作
* @param array
*/
def process(array:List[(TIME,THREAD,SERVICE,WAYS)]): Unit = {
val pendings = collection.mutable.Map[(THREAD,SERVICE),TIME]()
array.foreach{
case (time,thread,service,ways) => if(ways=="request") {
pendings += (thread,service) -> time
}
else {
pendings.get((thread,service)) match {
case Some(time0) => val msg = s"Thread: ${thread} Service: ${service} average-time: ${BigDecimal(sdf.parse(time).getTime - sdf.parse(time0).getTime)} ms"
pendings.remove((thread, service))
writeLogToFile(msg) ; println(msg)
case None => println("nothing")
}
}
}
}
/**
* 使用尾递归的方法，消除 var 和 mutable Map
* @param array
* @param pendings
*/
@tailrec
def processByTailRec(array:List[(TIME,THREAD,SERVICE,WAYS)],pendings:Map[(THREAD,SERVICE),TIME]): Unit ={
array match {
case (time,thread,service,"response") :: tail =>
pendings.get((thread, service)) match {
case Some(time0) =>
val msg = s"Thread: ${thread} Service: ${service} average-time: ${BigDecimal(sdf.parse(time).getTime - sdf.parse(time0).getTime)} ms"
writeLogToFile(msg) ; println(msg)
processByTailRec(tail, pendings - ((thread, service)))
case None =>
processByTailRec(tail, pendings)
}
case (time, thread, service, "request") :: tail =>
processByTailRec(tail, pendings + ((thread, service) -> time) )
case Nil => None
}
}
/**
* 写文件
* @param args
*/
def writeLogToFile(args:String): Unit ={
val fw = new FileWriter(new File("E:\\scala\\service_average.log"),true)
//写入中文字符时会出现乱码
val bw = new BufferedWriter(fw)
bw.write(args + "\r\n")
bw.close()
fw.close()
}
}