I am having trouble with the Linux find and grep commands. Basically I need to CD to a directory, recursive find all example.txt files modified after a particular date and then execute grep on those files. The command below is working (somewhat) but it gives multiple results of same file and I am not sure how stable it is.
find ./ -name "example.xml" -newermt "2018-01-01" -exec grep -r "videoplayer.1" /. \;
I would think the correct form of that command is:
find ./ -name "example.xml" -newermt "2018-01-01" -exec grep -r "videoplayer.1" {} \;
Missing the {} part in the find command. Your find results aren't being passed to grep.
\; and use + and see if it helps
\; you are passing only a single file at a time to grep, so it won't print the filename by default - you will need to add the -H option if you want it to add the name. Or add a -print to the find command.
+ not be ok too, and the -print prints all file names
find ./ -name "example.xml" -newermt "2018-01-01" -exec grep -r "videoplayer.1" /. \;
This would find, starting at the current directory ./, files with the -name example.xml that are newer (in modification time) than 2018-01-01 and for each of them, execute the command grep -r "videoplayer.1" /..
But we see that the grep command does not take the file name found by find into account, but instead the command is always constant. Furthermore, it starts a recursive (-r) search from the root of the filesystem (/.). Now, that may be a typo (./ vs. /.), but the meaning changes completely.
Even with grep -r ... ./, the search would be recursive starting from the current directory, and with -exec, that's the directory where find was started.
With find -exec, you need to use the {} placeholder to tell find where to insert the filename currently being processed. So
find ./ -name "example.xml" -newermt "2018-01-01" -exec grep "videoplayer.1" {} \;
To run grep videoplayer.1 filename on all files named example.xml that are newer than January 1st 2018.
Actually, you could do replace the \; that terminates the -exec command with a + to have find pass multiple files to one invocation of grep. That would make the whole job faster, since less processes need to be started:
find ./ -name "example.xml" -newermt "2018-01-01" -exec grep "videoplayer.1" {} \+
Also, if you want to make sure grep lists the name of the file it's processing, you can use the -H flag (in at least GNU and BSD grep):
-H, --with-filenamePrint the file name for each match. This is the default when there is more than one file to search.
(An alternative to -H would be to add something like /dev/null to the list of files given to grep)
-H option, which tells it to print the file name in any case. It's what happens by default if multiple files are listed. Also, I forgot to mention find -exec ... {} +, which would be better. Edited notes on both.
{}at the end of grep! Perhaps this would help:find ./ -name "example.xml" -newermt "2018-01-01" -exec grep -r "videoplayer.1" {} \;