Difference between revisions of "grep"

Latest revision as of 15:13, 21 June 2022

Pages

manpage - the current manpage
/manpage - a nicely-formatted but old version of the grep manpage

Examples

grep text_to_find *.log: find [text_to_find] by checking all the logfiles in the current folder
grep -r text_to_find *: find [text_to_find] by checking all files in the current folder or under it

Note that you can't do e.g. grep -r [text_to_find] *.log because that will only search in folders named [something].log. The reasons for this are complicated.

Questions

So, how do you grep for a string which contains spaces and quote characters? This is probably some standard feature of Linux Shell syntax which everyone knows, but I don't, and that's why this page needs examples.
A: Tentatively, you can surround a string with double-quotes and escape any quotes in the search-string with the backslash character. (Single quotes do not seem to work.)
Likewise, how do you search a group of files? Using "*.*" for FILE doesn't seem to work, and leaving the FILE argument blank tells grep to expect input from STDIN.
A: apparently "*" by itself; see examples above

Notes

The backtick character ("`") seems to mean "insert the output of this command here" -- so if you're trying to search for a literal backtick, it needs to be escaped with "\" ("\`"). This comes up sometimes when searching for SQL phrases.

Links

official documentation: more current than the manpage on this server
Wikipedia

@@ Line 18: / Line 18: @@
 *: '''A:''' apparently "*" by itself; see examples above
 ==Notes==
-If ''file'' is specified using wildcards (e.g. *.log or *), apparently the command-line parser expands this into a listing of all matching files before passing it to grep. This unfortunately can lead to the error "bash: /bin/grep: Argument list too long". Putting a backslash before the wildcard will prevent the parser from expanding it before passing to grep, but grep doesn't seem to have any wildcard-handling abilities. So as far as I can tell, there is no way to grep a really large directory (the one I'm looking at has 5014 files).
+The backtick character ("`") seems to mean "insert the output of this command here" -- so if you're trying to search for a literal backtick, it needs to be escaped with "\" ("\`"). This comes up sometimes when searching for [[SQL]] phrases.
-The -r option doesn't seem to work if a file extension is specified (e.g. "*.log" in the example above), though it does work with "*". There may be some syntax which needs to surround "*.log* -- or maybe this simply is not doable using the inline-expansion method of handling wildcards?
-The backtick character ("`") seems to mean "insert the output of this command here" -- so if you're trying to search for a literal backtick, it needs to be escaped with "\" ("\`"). This comes up sometimes when searching for [[SQL]] phrases.
 ==Links==
 * [http://www.gnu.org/software/grep/manual/ official documentation]: more current than the manpage on this server
 * {{wikipedia}}