Bash grep regex问题有两个不同的文件

Question

I have the following command which is filtering 3-letters words from a file made of upper case words only - one word per line:

grep -E '^[A-Z]{3}$' test

The command returns a correct list of words when used with a file test containing 10 words. When applied to a much bigger file dico.txt containing over 30,000 words, the command does not return anything (a new prompt is simply displayed).

As I thought it might be either an extension or a file size issue, I've tried:

• cp test test.txt to match the big file *.txt extension
• Create a new file dico_small.txt selecting 1000 lines from dico.txt

...both without success.

Answer 1

Your large file has windows line endings, that is \\r\\n instead of linux line endings \\n .

\\r is called carriage return and is treated as a normal character by grep . When you write grep -E "a$" fileWithWindowsLineEndings then grep won't find anything because in front of the linux line ending \\n (denoted as $ in grep) there is always a \\r and never an a .

You can convert your file to a normal linux file by deleting all \\r characters.

tr -d '\r' < fileWithWindowsLineEndings > fileWithLinuxLineEndings
grep -E '...' fileWithLinuxLineEndings

Alternatively, convert the file on the fly without saving the conversion result:

tr -d '\r' < fileWithWindowsLineEndings | grep -E '...'