-3

我有一个像下面的文件

ElapsedTime2.68s: PlaceOrder
ElapsedTime2.69s: ClassARestCAll
ElapsedTime0.11s: GetOrderList
ElapsedTime0.11s: ClassARestCAll
ElapsedTime2.10s: PlaceOrder
ElapsedTime2.11s: ClassARestCAll
ElapsedTime0.10s: GetOrderList
ElapsedTime0.10s: ClassARestCAll
ElapsedTime2.00s: PlaceOrder
ElapsedTime2.01s: ClassARestCAll
ElapsedTime0.28s: GetOrderList
ElapsedTime0.28s: ClassARestCAll
ElapsedTime1.64s: PlaceOrder
ElapsedTime1.65s: ClassARestCAll
ElapsedTime0.11s: GetOrderList
ElapsedTime0.11s: ClassARestCAll
ElapsedTime1.99s: PlaceOrder
ElapsedTime2.01s: ClassARestCAll

如何解析文件以获得此结果？

               average   min   max
ClassARestCAll    1.23   0.1  2.69
GetOrderList      0.15   0.1  0.28
PlaceOrder       2.082  1.64  2.68

我开发了一种使用 RE 和 List 操作来解决它的方法。但是，我的方法为每个新方法名称扫描整个列表一次。

我们如何通过仅扫描一次列表来获取所有 API 名称的统计信息。

import re

def get_stats(N, p_api):
    list_of_rt = []
    for line in N:
        y= re.split("\s+", line)
        if y[1] == p_api:
             curr_rt = float(y[0][11:-2])
             list_of_rt.append(curr_rt)

    min_rt ,max_rt = min(list_of_rt), max(list_of_rt)
    total_rt, total_cnt = sum(list_of_rt), len(list_of_rt)
    print p_api, min_rt, max_rt, "%.3f" %round(total_rt/total_cnt,3), total_cnt


ifile = open('data1.txt','r').read()
api_rts= re.findall(r'ElapsedTime\d*.\d*s: \S*',ifile)


list_of_api_names = []
for api_rt in api_rts:
    y= re.split("\s+", api_rt)
    list_of_api_names.append(y[1])

#get distinct list of API names

distinct_apis = set(list_of_api_names)

print 'api         min,   max,   average,  total occurences'

# for each API name call get_stat

for api in distinct_apis:
    get_stats(api_rts ,api)
4

2 回答 2

1 
import re

rgx = re.compile('ElapsedTime(\d*\.\d*)s: (\S*)')

from collections import defaultdict
d = defaultdict(list)

with open('data1.txt','r') as f:
    for m in rgx.finditer(f.read()):
        d[m.group(2)].append(float(m.group(1)))

lapi = max(map(len,d.iterkeys()))

print '{: ^{width}}   min    max    average   total occurences'.format('api',width=lapi)
pat = '{0:%d}   {1:.2f}   {2:.2f}   {3:.3f}     {4}' % lapi
print '\n'.join(pat.format(api,min(li),max(li),sum(li)/len(li),len(li))
                for api,li in d.iteritems())
于 2013-08-26T08:50:29.827 回答
0 
import re, numpy
from collections import defaultdict

data = defaultdict(list)
with open('data1.txt') as f:
    for line in f:
        l =  re.findall('([\d.]+|\w+$)', line)
        data[l[1]].append(float(l[0]))

metrics = [ ['avg', numpy.average], ['min',min], ['max',max] ]
summary = defaultdict(dict)
for k,l in data.items():
    summary[k] = { m[0] : m[1](l) for m in metrics }

print " " * 17 + "\t".join("%-5s" % m[0] for m in metrics)
for k, s in summary.items():
    print "%-15s: " % k + "\t".join("%.3f" % s[m[0]] for m in metrics)
于 2013-08-26T06:35:34.207 回答