0

考虑代码:

我在哪里读取带有 6 列(0-5)的输入文件

  1. 将变量 historyends 初始化为 5000。
  2. 接下来,当 column0 值 i,e job[0] < 5000 时,我将输入文件的 5000 行添加到一个列表(historyjobs)中,否则其余行直到另一个列表中的 eof(targetjobs)。
  3. 接下来所有的historyjobs 列出item3,item4,item5 中的所有内容等于targetjobs first list item3,item4,item5 当这个条件满足时将那些historyjobs 所有item1 添加到listsub。
  4. 接下来找到listsub中项目的运行平均值并反转列表,将其存储在列表a中。如果listsub中的项目> a * 0.9,则检查条件满足条件将结果项目存储在列表condsub中。
  5. 接下来重新打开输入文件并检查 column0 是否等于 condsub 中的项目,如果满足则将 column1 添加到列表 condrun。
  6. 最后打开输出文件并在colum0中写入targetjobs中第一个列表的第二项i,ej,在column1中写入列表condrun的平均值,column2是(j-avg)/j,column3是list condrun中的最大项目,column4是列表 condrun 中的最小项目,第 5 列是列表 condrun 的长度,最后四列是基于条件的。
  7. 最后,我通过将变量 historyends 分配给下一项 int(targetjobs[1][0]) 使用 while 循环重复整个过程

    from __future__ 
    import division
    import itertools   
    history_begins = 1; history_ends = 5000; n = 0; total = 0
    historyjobs = []; targetjobs = []
    listsub = []; listrun = []; listavg = [] ; F = [] ; condsub = [] ;condrun = [] ;mlistsub = []; a = []
    
    def check(inputfile):
    
      f = open(inputfile,'r') #reads the inputfile
      lines = f.readlines()
      for line in lines:
          job = line.split()
          if( int(job[0]) < history_ends ): #if the column0 is less then history_ends(i,e 5000 initially)
                historyjobs.append(job) #historyjobs list contains all the lines from the list whose column1 < history_ends
          else:
                targetjobs.append(job) #historyjobs list contains all the lines from the list whose column1 > history_ends  
      k = 0           
      for i, element in enumerate(historyjobs):
          if( (int(historyjobs[i][3]) == int(targetjobs[k][3])) and (int(historyjobs[i][4]) == int(targetjobs[k][4])) and (int(historyjobs[i][5]) == int(targetjobs[k][5])) ): #historyjobs list all contents in column3,column4,column5 is equal to targetjobs first list column3,column4,column5
    
                 listsub.append(historyjobs[i][1]) #when if condition true add those historyjobs column1 to list listsub
    
    def runningMean(iterable):
    """A generator, yielding a cumulative average of its input."""
      num = 0
      denom = 0
      for x in iterable:
    num += x
    denom += 1
    yield num / denom
    
    def newfun(results):
      results.reverse() # put them back in regular order
      for value, average in results:
    a.append(value)
      return a #to return the value   
    
    
    def runcheck(subseq):
      f = open('newfileinput','r') #again read the same inputfile
      lines = f.readlines()
      for line in lines:
          job = line.split()
          for i, element in enumerate(subseq):
         if(int(job[1]) == int(subseq[i])): # if the column1 value of the inputfile becomes equal to list obtained
              condrun.append(str(job[2])) #return the value of column2 which satisfies the if condition
      return condrun
    
    
    def listcreate(condrun,condsub):
     f1 = open('outputfile','a') #outputfile to append the result
     s = map(int,condrun)
     j = int(targetjobs[0][2])
     targetsub = int(targetjobs[0][1])
     if(condsub != []):
      try:
       convertsub = int(condsub[-1])
       a=sum(s)/len(s)
       c=max(s)
       d=min(s)
       e1=abs(j-a)
       er1=e1/j
       g=len(s)
       h=abs(convertsub-targetsub)
       f1.write(str(j))
       f1.write('\t')
       f1.write('\t')
       f1.write(str(round(a,2)))
       f1.write('\t')
       f1.write('\t')
       f1.write(str(round(er1,3)))
       f1.write('\t')
       f1.write('\t')
       f1.write(str(c))
       f1.write('\t')
       f1.write('\t')
       f1.write(str(d))
       f1.write('\t')
       f1.write('\t')
       f1.write(str(g))
       f1.write('\t')
       f1.write('\t')
       f1.write(str(h))
       f1.write('\t')
       f1.write("\t")
       if (float(er1) < 0.20):
         f1.write("good")
         f1.write("\t")
       else :
         f1.write("bad")
         f1.write("\t")
       if (float(er1) < 0.30):
         f1.write("good")
         f1.write("\t")
       else :
         f1.write("bad")
         f1.write("\t")
       if (float(er1) < 0.40):
         f1.write("good")
         f1.write("\t")
       else :
         f1.write("bad")
         f1.write("\t")
       if (float(er1) < 0.50):
         f1.write("good")
         f1.write("\n")
       else :
         f1.write("bad")
         f1.write("\n")
      except ZeroDivisionError :
       print 'dem 0'
     else:
       print '0'
       f1.close() 
    
    def new():
     global history_ends
     while 1: #To repeat the process untill the EOF(end of input file)
    
    check('newfileinput') #First function call
    if(len(targetjobs) != 1):
     history_ends = int(targetjobs[1][0]) #initialize historyends to targetjobs second lines first item
     mlistsub = map(int,listsub)
     results = list(itertools.takewhile(lambda x: x[0] > 0.9 * x[1],
                   itertools.izip(reversed(mlistsub),
                          runningMean(reversed(mlistsub)))))#call runningmean function & check the condition
     condsub = newfun(results) #function to reverse back the result    
     condrun=runcheck(condsub) #functionto match & return the value
     listcreate(condrun,condsub) #function to write result to output file     
     del condrun[0:len(condrun)]#to delete the values in list
     del condsub[0:len(condsub)]#to delete the values in list
     del listsub[0:len(listsub)]#to delete the values in list
     del targetjobs[0:len(targetjobs)]#to delete the values in list
     del historyjobs[0:len(historyjobs)]#to delete the values in list
    else:
      break
    
    
    def main():      
       new()
    
    if __name__ == '__main__':
    main()
    

示例输入文件(整个文件包含 200,000 行):

 1  0   9227    1152    34  2
 2  111 7622    1120    34  2
 3  68486   710 1024    14  2
 6  265065  3389    800 22  2
 7  393152  48438   64  132 3
 8  412251  46744   64  132 3
 9  430593  50866   256 95  4
 10 430730  10770   256 95  4
 11 433750  12701   256 14  3
 12 437926  2794    64  34  2
 13 440070  43  32  96  3
 13 440070  43  32  96  3
 14 440102  44  32  96  3
 15 440357  43  32  96  3
 16 440545  43  32  96  3
 17 440599  43  32  96  3
 18 440625  43  32  96  3
 19 440999  84  32  96  0
 20 441574  44  32  96  3
 21 442667  7914    512 14  3
 22 443249  45  32  96  3
 25 443797  3260    128 68  4
 26 443799  3746    128 68  4
 27 445357  31  8   29  3
 28 445393  31  8   29  3
 29 445445  28  8   29  3
 30 445488  29  8   29  3

示例输出文件内容:

    930     1389.14     0.494       3625        977     7       15      bad bad bad good
    4348        1331.75     0.694       3625        930     8       164     bad bad bad bad
    18047       32237.0     0.786       61465       17285       3       325774      bad bad bad bad
    1607        1509.0      0.061       1509        1509        1       6508        good    good    good    good
    304     40.06       0.868       80      32      35      53472       bad bad bad bad
    7246        7247.0      0.0     7247        7247        1       9691        good    good    good    good
    95      1558.0      15.4        1607        1509        2       2148        bad bad bad bad
    55      54.33       0.012       56      53      3       448142      good    good    good    good
    31      76.38       1.464       392     35      13      237152      bad bad bad bad
    207     55.0        0.734       55      55      1       370     bad bad bad bad

如果有人可以建议一些更改以使代码运行得更快,那将很有帮助...

输入文件:

  column 0-->represents jobnum
  column 1-->represents submittime
  column 2-->represents runtime
  column 3-->represents userid
  column 4-->represents numberof processor
  column 5-->represents queueid

我维护输入文件中的前 5000 行,因为某种历史读取第 5001 行比较其 col3、col4、col5 值是否等于 5000 行中的值。如果大约 20 行具有匹配值,则所有满足条件 col2 的行到一个列表。找到此列表的运行平均值并将结果存储到另一个列表 1。现在检查所有项目的条件(列表 1 > 列表 * 0.9)满足条件的项目将其添加到另一个列表列表 3。list3 中与历史记录中 5000 行的 col0 匹配的所有项目,将 col2 存储到 list4。现在我需要打开一个文件来写入最终 list4 的平均值。重复相同的操作,将历史记录增加到下一行,直到 EOF

4

2 回答 2

3

两者都check具有runcheck读取整个文件的代码,然后遍历所有行。替换以下内容:

lines = f.readlines()
for line in lines:

和:

for line in f:

一次只读取和处理行。

于 2013-08-31T15:25:03.367 回答
1

我认为以下代码与您的代码相同:

from __future__ import division
from sys import exit
history_begins = 1
history_ends = 16
historyjobs = []
targetjobs = []

def quickzh(zhlistsub,
            historyjobs=historyjobs):
    rev = reversed(zhlistsub)
    i = next(rev)
    num   =  historyjobs[i][1]
    denom = 1
    hfirst  = num
    li = [historyjobs[i][2]]
    for i in rev:
        x = historyjobs[i][1]
        num += x
        denom += 1
        if x > 0.9 * (num / denom):
            li.append(historyjobs[i][2])
        else:
            break
    li.reverse()
    return hfirst,li 


def listcreate(hfirst,s,
               historyjobs=historyjobs,targetjobs=targetjobs):
    with open('outputfile.txt','a') as f1:
        j = targetjobs[0][2]
        try:
            a,c,d,g = sum(s)/len(s), max(s), min(s), len(s)
            e1  = abs(j-a)
            er1 = e1/j
            h   = abs(hfirst-targetjobs[0][1])
            # historyjobs[-1][1] is convertsub
            # targetjobs[0][1]   is targetsub
            f1.write('%s\t%s\t%s\t%s\t%s\t%s\t%s\t' %
                     (j, round(a,2), round(er1,3), c, d, g, h))
            f1.write("%s\t%s\t%s\t%s\n" %
                     tuple("good" if er1 < x else "bad" for x in (0.2,0.3,0.4,0.5)))
        except ZeroDivisionError :
            print 'dem 0'
        else:
            print '0'
            f1.close()

def new(inputfile,history_ends,
        historyjobs=historyjobs,targetjobs=targetjobs):
    while 1:

        # checking the file
        with open(inputfile,'r') as f:
            for line in f:
                job = map(int,line.split())
                (historyjobs if job[0]  < history_ends
                 else targetjobs).append(job)  

        if len(targetjobs) != 1:
            k = 0           
            zhlistsub = [i for i, element in enumerate(historyjobs)
                         if element[3:6] == targetjobs[k][3:6] ]
            if zhlistsub:
                listcreate(*quickzh(zhlistsub))
            history_ends = targetjobs[1][0]
            del targetjobs[:]
            del historyjobs[:]
        else:
            break


new('toto.txt',history_ends)

使用您提供的输入样本 和history_ends = 16,输出文件变为:

43  43.5    0.012   44  43  2   188 good    good    good    good
43  43.33   0.008   44  43  3   54  good    good    good    good
43  43.25   0.006   44  43  4   26  good    good    good    good
44  43.2    0.018   44  43  5   949 good    good    good    good
45  49.14   0.092   84  43  7   1675    good    good    good    good
3746    3260.0  0.13    3260    3260    1   2   good    good    good    good
31  31.0    0.0 31  31  1   36  good    good    good    good
28  31.0    0.107   31  31  2   52  good    good    good    good

如果你需要解释,问我。
原则是跟踪指数,而不是历史工作中的值。

只是备注:
在生成器函数中runningMean()

for x in iterable:
    num += x
    denom += 1
    yield num / denom

对于可迭代的第一个元素,denom等于 1 然后num/denom等于num
因此,在takewhile(lambda x: x[0] > 0.9 * x[1], izip(reversed(mlistsub), runningMean(reversed(mlistsub)))))
第一个元素x中始终是 kind (el,el), thenx[0] > 0.9 * x[1]对于第一个元素始终为 true ,然后resultsandconsub永远不会是 void 列表。
所以在listcreate()函数中,condsub传递的参数永远不会为空,并且条件if consub != |]始终为真。
这就是为什么在我的代码中,这种情况已经消失了。

.

编辑

如果输入文件的第一列的值(读取为 job[0])是递增值,您可以修改new()为:

def new(inputfile,history_ends,
        historyjobs=historyjobs,targetjobs=targetjobs):

    # checking the file
    with open(inputfile,'r') as f:
        for line in f:
            job = map(int,line.split())
            (historyjobs if job[0]  < history_ends
             else targetjobs).append(job)

    while True:
        k = 0           
        zhlistsub = [i for i, element in enumerate(historyjobs)
                     if element[3:6] == targetjobs[k][3:6] ]
        if zhlistsub:
            listcreate(*quickzh(zhlistsub))
        tj00 = targetjobs[0][0]
        while True:
            if targetjobs[0][0]!=tj00:
                break
            historyjobs.append(targetjobs.pop(0))
        if len(targetjobs)==0:
            break

如果值严格增加,也就是说在第 1 列中没有两行具有相同的值,我认为您可以简化为:

def new(inputfile,history_ends,
        historyjobs=historyjobs,targetjobs=targetjobs):

    # checking the file
    with open(inputfile,'r') as f:
        for line in f:
            job = map(int,line.split())
            (historyjobs if job[0]  < history_ends
             else targetjobs).append(job)

    while True:
        k = 0           
        zhlistsub = [i for i, element in enumerate(historyjobs)
                     if element[3:6] == targetjobs[k][3:6] ]
        if zhlistsub:
            listcreate(*quickzh(zhlistsub))
        historyjobs.append(targetjobs.pop(0))
        if len(targetjobs)==0:
            break

注意我仍然不明白需要k

于 2013-09-01T17:57:11.753 回答