实验需要,在一个含有几亿个字符的txt文件中查找特定的字符串,首先用re模块进行查找
from time import clock
import re
start=clock()
label_file = open("/home/ying/data/google_streetview_train_test1/label.txt")
label_str = label_file.read()
label_file.close()
filename = "2_0_pitch_95_yaw_95_lat_41.8975137_lng_-87.6268723.jpg"
start=clock()
for match in re.finditer(filename, label_str):
s = match.start()
e = match.end()
print(s)
print(e)
end=clock()
print(end-start)
re.finditer(filename, label_str)可以在label_str中查找filename的位置,s=match.start()返回字符串开始的索引,e=match.end(),返回字符串结束的索引。程序运行的结果是
304091635
304091689
304096479
304096533
1.003844
耗时1s左右
同样的,由于txt文件中为一行一行的数据,可以用readlines进行遍历读取比较,程序如下
from time import clock
start=clock()
data_label="/home/ying/data/google_streetview_train_test1/label.txt"
filename = "2_0_pitch_95_yaw_95_lat_41.8975137_lng_-87.6268723.jpg"
file = open(data_label)
lines = file.readlines()
print(len(lines))
for line in lines:
cls = line.split()
fn = cls.pop(0)
if fn==filename:
break
end=clock()
print(end-start)
运行结果如下:
1
3.335657
可见耗时有3s多,用正则化模块要快的多
另外,由于label_str中存在1.2_0_pitch_95_yaw_95_lat_41.8975137_lng_-87.6268723.jpg,所以用re模块寻找时会返回两个结果,而用逐行读取的方式则返回一个值