How to search backward several lines in Python 3?

Question

There is a solution for searching backward inline in Python Reverse Find in String:

s.rfind('I', 0, index)

But if I need to search for a string in several lines above that line? Say I have found the keyword by using:

with open("file.txt") as f
    searchlines = f.readlines()

for i, line in enumerate(searchlines):
    if "keyword" in line: 
    do_something()

I want do_something() is to find another keyword backward. To apply the code above, I think I need to f.read() so that I can make the file as a string. But this is totally nut, since I have to readlines() and read() the (large) file. I need to use readlines() because the first keyword may appears several times in the text, and I need to find them all.

Is there any better way to do this?

image description

@engineer
- kỹ sư
@engineering
- kỹ thuật
- civil e. ngành xây dựng
- communication e. kỹ thuật thông tin
- control e. kỹ thuật [điều chỉnh, điều khiển] (tự động)
- development e. nghiên cứu những kết cấu mới

could you explain the pic little bit? And it would be nice if those are characters rather than an image. — Avinash Raj, Aug 30 '15 at 14:01
Which part you don't understand? Doesn't my question explain the pic already? I don't know if you need the lines from the pic or not. It is there just for illustration. But if you insist... — Ooker, Aug 30 '15 at 14:04
This question could be useful: [How to read from a file in python starting from the end](http://stackoverflow.com/questions/3568833/how-to-read-lines-from-a-file-in-python-starting-from-the-end) — Mel, Aug 30 '15 at 15:12

score 4 · Accepted Answer · answered Aug 30 '15 at 15:25

I'd rather approach this this way: since you want to find the line starting with @, I'd rather store all the lines in a list, then discard the previous lines if a new line that starts with @ is found.

Thus we get:

def do_something(lines):
    print("I've got:")
    print(''.join(lines))

lines = []

with open("file.txt") as f:
    for i, line in enumerate(f):
        if line.startswith('@'):
            lines = []

        lines.append(line)
        if 'development' in line:
            do_something(lines)

The output with file.txt as you have, will be:

I've got:
@engineering
- kỹ thuật
- civil e. ngành xây dựng
- communication e. kỹ thuật thông tin
- control e. kỹ thuật [điều chỉnh, điều khiển] (tự động)
- development e. nghiên cứu những kết cấu mới

In general case if you want to have just N last seen lines, you can use a collections.deque instead of a list:

from collections import deque
N = 100
last_lines = deque(maxlen=N)

with open("file.txt") as f:
    for i, line in enumerate(f):
        last_lines.append(line)
        if 'development' in line:
            do_something(last_lines)

Now the do_something will be passed up to 100 last lines including the current line, if the current line contains the word development.

Simply brilliant. Instead of searching backward, you just remember all the text between two keywords and reset the "memory" if it is not the right one. — Ooker, Aug 30 '15 at 18:52

How to search backward several lines in Python 3?

1 Answers1