How to get list of file names and their sentence containing a specific search string?

Question

I tried with following command on unix machine:

ls -l | awk '{print $9}' | xargs -I {} cat  {}  | grep {"String to search"}

Though this works with text files but when i try it with xml files it is not able to display proper grepped text.Instead it displays whole xml file.

I think the possible reason behind this is absense of new line character in xml file that i use.

Example: Search string: "/1031/"

Xml line containing search string: <eventtype uri="{any_url}/1031/"/>

To clarify a bit :

ls -l | awk '{print $9}' | xargs -I {} cat {} | grep -o "/1031"

This gives output as:

/1031

/1031

/1031...

I also want the name of the file in which it belongs.

Answer 1

grep has a flag -o which only outputs the matching text.

ls -l | awk '{print $9}' | xargs -I {} cat {} | grep -o {"String to search"}

From your edit it looks like you need the "line" that contains the URL as well. By default grep will match greedily which means a regex to account for the XML formatting will still give you an incorrect result.

I can think of 2 possible options:

For the next examples, test.xml contains the string:

<eventtype uri="{www.example1.com}/1031/"/><eventtype uri="{www.example2.com}/1031/"/><eventtype uri="{www.example3.com}/1031/"/>

The first is to use the -P flag for grep to enable perl syntax and match lazily.

grep -Po '".*?/1031/"' test.xml 

This outputs:

"{www.example1.com}/1031/"
"{www.example2.com}/1031/"
"{www.example3.com}/1031/"

The second is to use sed to manually append a newline after each match and pipe to grep:

sed 's/1031/1031\n/g' test.xml | grep 1031

Outputs:

<eventtype uri="{www.example1.com}/1031
/"/><eventtype uri="{www.example2.com}/1031
/"/><eventtype uri="{www.example3.com}/1031

I believe both methods should work ok on plain text files although you may need to conditionally use one of these methods on .xml extensions.

Answer 2

Dear You could use below command

find -type f -exec grep -HPo '".*?/1031/"' {} \;

Sample Output

[root@MUM03S001 ~]# find -type f -exec grep -HPo '".*?/1031/"' {} \\;

./File:"{www.example1.com}/1031/"

./File:"{www.example2.com}/1031/"

./File:"{www.example3.com}/1031/"

[root@MUM03S001 ~]#