Odtnhj

I have a txt-file with sentances and am able to find words from a list within it. I would like to print the line above the 'found-line' to a seperate list. I tried it with the below-code, but this only returns .

Here is my code:

fname_in = "test.txt"
lv_pos = 
search_list = ['word1', 'word2']

with open (fname_in, 'r') as f:
 file_l1 = [line.split('n') for line in f.readlines()]
 counter = 0

 for word in search_list:
 if word in file_l1:
 l_pos.append(file_l1[counter - 1])

 counter += 1

print(l_pos)

The text file looks somthing like this:

Bla bla bla
I want this line1.
I found this line with word1.
Bla bla bla
I want this line2.
I found this line with word2.

The result I want this:

l_pos = ['I want this line1.','I want this line2.']

In the second line of your example you wrote lv_pos instead of l_pos. Inside the with statement you could fix it like this I think:

fname_in = "test.txt"
l_pos = 
search_list = ['word1', 'word2']

file_l1 = f.readlines()

for line in range(len(file_l1)):
 for word in search_words:
 if word in file_l1[line].split(" "):
 l_pos.append(file_l1[line - 1])

print(l_pos)

I'm not thrilled about this solution but I think it would fix your code with minimal modification.

Treat the file as a collection of pairs of lines and lines-before:

[prev for prev,this in zip(lines, lines[1:]) 
 if 'word1' in this or 'word2' in this]
#['I want this line1.', 'I want this line2.']

This approach can be extended to cover any number of words:

words = 'word1', 'word2'
[prev for prev,this in zip(lines,lines[1:]) 
 if any(word in this for word in words)]
#['I want this line1.', 'I want this line2.']

Finally, if you care about proper words rather than occurrences (as in "thisisnotword1"), you should properly tokenize lines with, say, nltk.word_tokenize():

from nltk import word_tokenize
[prev for prev,this in zip(lines,lines[1:]) 
 if words & set(word_tokenize(this))]
#['I want this line1.', 'I want this line2.']

First of all you got some typos in your code—in some places you wrote l_pos and in others, lv_pos.

The other problem is I don't think you realize that file_l1 is a list-of-lists, so the if word in file_l1: isn't doing what you think. You need to check each word against each of these sublists.

Here's some working code based on your own:

fname_in = "simple_test.txt"
l_pos = 
search_list = ['word1', 'word2']

with open(fname_in) as f:
 lines = f.read().splitlines()

 for i, line in enumerate(lines):
 for word in search_list:
 if word in line:
 l_pos.append(lines[i - 1])

print(l_pos) # -> ['I want this line1.', 'I want this line2.']

Update

Here's another way to do it that doesn't require reading the entire file into memory at once, so doesn't require as much memory:

from collections import deque

fname_in = "simple_test.txt"
l_pos = 
search_list = ['word1', 'word2']

with open(fname_in) as file:
 lines = (line.rstrip('n') for line in file) # Generator expression.

 try: # Create and initialize a sliding window.
 sw = deque(next(lines), maxlen=2)
 except StopIteration: # File with less than 1 line.
 pass

 for line in lines:
 sw.append(line)
 for word in search_list:
 if word in sw[1]:
 l_pos.append(sw[0])

print(l_pos) # -> ['I want this line1.', 'I want this line2.']