I'm trying to read through a file, find a certain pattern and then grabbing a set number of lines of text after the line that contains that pattern. Not really sure how to approach this.
+1
A:
First parse the file into lines. Open, read, split on the line break
lines = File.open(file_name).read.split("\n")
Then get index
index = line.index{|x| x.match(/regex_pattern/)}
Where regex_pattern is the pattern that you are looking for. Use the index as a starting point and then the second argument is the number of lines (in this case 5)
lines[index, 5]
It will return an array of 'lines'
You could combine it a bit more to reduce the number of lines. but I was attempting to keep it readable.
Geoff Lanotte
2010-07-11 04:59:29
Since this opens up the entire file at once would it have memory problems at all with bigger documents? (up to 5mb or so)
Randuin
2010-07-11 08:05:39
not so much, I use similar methods to parse log files of 500MB.
Geoff Lanotte
2010-07-11 11:30:42
Sucking the entire file into memory is poor style, even though VM enables it.
Larry K
2010-07-11 17:47:14
What would be the correct style to maintain functionality but in "good style"?
Randuin
2010-07-11 18:20:21
lars has the answer, give it to him.
Geoff Lanotte
2010-07-11 20:11:01
A:
matched = false;
num = 0;
res = "";
new File(filename).each_line { |line|
if (matched) {
res += line+"\n";
num++;
if (num == num_lines_desired) {
break;
}
} elsif (line.match(/regex/)) {
matched = true;
}
}
This has the advantage of not needing to read the whole file in the event of a match.
When done, res will hold the desired lines.
Borealid
2010-07-11 05:15:12
Holy moly. This has to be hands down the ugliest Ruby code I've ever seen. Well, actually, given the amount of syntax errors, I'd hardly even call it Ruby code.
Jörg W Mittag
2010-07-11 08:35:20
It's CRuby! Hey, Ruby supports block syntax. You don't *have* to use "begin" and "end" :-P
Borealid
2010-07-11 08:59:56
+1
A:
If you're not tied to Ruby, grep -A 12 trivet will show the 12 lines after any line with trivet in it. Any regex will work in place of "trivet"
Slartibartfast
2010-07-11 05:20:49
Well I actually need to still parse the following whatever lines. And extract information out of them. So grep wasn't a good fit.
Randuin
2010-07-11 05:36:18
+1
A:
If you want the n number of lines after the line matching pattern in the file filename:
lines = File.open(filename) do |file|
line = file.readline until line =~ /pattern/ || file.eof;
file.eof ? nil : (1..n).map { file.eof ? nil : file.readline }.compact
end
This should handle all cases, like the pattern not present in the file (returns nil) or there being less than n lines after the matching lines (the resulting array containing the last lines of the file.)
Lars Haugseth
2010-07-11 19:38:36