Mission impossible?

[Langstracht]

I have a bunch of files in a bunch of sub-directories which may or may not contain specific text.

For instance D001, D002, D003 ... may be in the files.

I need to output a list of which TEXT(s) exist. I don't want to know which file they are in only that they exist.

So I would look for "D0" in the directory and the output would be D001, D002, D003 etc.

I have fiddled around with find and grep to no avail.

Can anyone help?

Much appreciation in anticipation

[The Cog]

I assume they are all D followed by three digits. If so, maybe this would do the trick?

grep -R -o 'D[0-9][0-9][0-9]' | sort -u

[Langstracht]

Thanks for your prompt reply.

I guess I misled you. Sorry.

The sought text could be any length (D0010, D00022, D00045, ...) BUT all starting with D00 ...

So think:

look for "star", output start, started, starlight ...

[dragonfly41]

I suggest https://itsfoss.com/ripgrep-all/

using wildcards

also a GUI: Recoll (but RipGrep is easier for me)

[MAFoElffen]

I assume they are all D followed by three digits. If so, maybe this would do the trick?

grep -R -o 'D[0-9][0-9][0-9]' | sort -u

The sought text could be any length (D0010, D00022, D00045, ...) BUT all starting with D00 ...

Modified:

grep -E 'D00*[0-9][0-9][0-9]' ./* | sort -u

Note: Only one of 'very many ways'.

[The Cog]

Both of these would match the given string definition. It's a very flexible definition we've been given.

grep -E -r -o 'D00.+' | sort -u
grep -E -r -o 'D00[0-9]+' | sort -u

[Langstracht]

Your suggested commands "work" - thank you - but the output is not what I want/need.

As I tried to explain earlier, I need the output to be just the sought after text - D00111 or whatever. Not the complete line on which it occurs.

[The Cog]

You haven't defined the "sought after text". It seems to start with "D00" but "could be any length". Actually, with your current definition, everything from "D00" to the end of the file is a valid result.

[Langstracht]

I'm afraid I don't really understand the problem that you are having.

But let me try to do this again.

Any file in a folder may,or may not, contain the words star, startle, started, starlight.

I would like to discover which of these words are in a file and output them - the words NOT the line they appear on.

So I would command

grep -r "star"

and this WILL locate and output the line(s) on which any of these words appear.

So how do I get it to just output the found words?

[The Cog]

Define what constitutes the start and end of the search text. With words, we can perhaps assume that a non-alpha character (space, comma etc) marks the end of the word. With egrep, we could search for '\bstar\w*\b' except that grep thinks that \w (meaning letters in a word) includes underscore, so that stars_and_stripes would be seen as one word. That may or may not matter to you. Would "superstar" be a match for you?

But you have described wanting to search for "D00" followed by an indeterminate length something. Where does it start or end? Messing around, I found a file here that contains "MAP_RESERVED0080". Does that match? What about "D0080x994-w pop"?

This will find all strings starting with D00 followed by any numbers of digits. If it sees "MAP_RESERVED0080xyz" it will output "D0080".

grep -r -E -o -h 'D00[0-9]+' | sort -u