87

In bash, if I want to execute a command and only display output lines that matches a certain pattern, I can pipe it to grep, like

file testfile

hello
there
my
friends

command

$ cat testfile | grep 'hello'
hello #this will be highlightd

this will highlight the search match and display the entire line it falls on. I can use -A and -B to display lines before and after that line. My question is is it possible to execute the command and display all output as normal, but to highlight the search matches like grep would? so my ouput would be

hello #highlighted
there
my
friends
1

8 Answers 8

90

To use a Color GREP to only highlight matched patterns but not otherwise change the output:

grep --color=always -e "^" -e "hello" testfile

The first pattern will match all lines (all lines will be printed) the second pattern (and any following patterns) cause the matched text to be highlighted in color.

Since the first pattern matches all lines but does not match on a printable character, it does not add any color highlighting so it doesn't compete/interfere with the readability of the highlighted text.

6
  • Works with kubectl kubectl apply -k . | grep --color=always -e "^" -e "configured" in zsh Jun 3, 2019 at 13:27
  • 1
    For use in Bash: highlight() { grep --color=always -e "^" -e $* ; }
    – Jon
    Nov 20, 2019 at 13:56
  • This works on FreeBSD as well, you just need to swap the first and the second patern ("^" needs to be second). Nov 22, 2019 at 6:57
  • Very smart! Thanks! Have my upvote! Jun 4, 2021 at 7:37
  • 1
    I like this version: egrep --color "^|hello" testfile
    – Alek
    Dec 7, 2021 at 22:41
40

Add option -z to your GNU grep command:

cat testfile | grep --color=always -z 'hello'

or shorter

grep --color=always -z 'hello' testfile
7
  • 6
    This is a very crude hack.   It has the effect of treating the entire file as a single line.   Therefore, (1) if the file is very large, there may be a possibility of running out of memory, and (2) if the file doesn’t contain the pattern at all, then nothing will be output. Feb 6, 2017 at 4:41
  • What version of grep is this supported on? On grep 2.5.4, -z doesn't seem available...
    – Alex
    Jun 29, 2017 at 22:29
  • @Alex: I used GNU grep version 2.6.3.
    – Cyrus
    Jun 30, 2017 at 2:37
  • 3
    @G-Man: more resource-saving: grep --color "hello\|$" file
    – Cyrus
    Feb 9, 2018 at 17:06
  • @Cyrus: Yes, that’s a better answer. Unfortunately, it is essentially equivalent to bot779’s answer to this question, and almost identical to the accepted answers (1 and 2) to the questions I linked to. Feb 9, 2018 at 18:10
10

This one works with GNU grep as well as with grep on FreeBSD:

grep --color=always 'hello\|$'

It matches the text "hello" or (\|) the non-printable null string at the end of each line ($). That's why each line gets printed but only "hello" is highlighted.

Chances are you already have --color=auto configured in your shell. Then you most likely don't need to specify --color=always:

grep 'hello\|$'

You can also simpler version by using egrep (with extended regular expressions), where | for the "or" expression doesn't need to be escaped:

egrep 'hello|$'

See also similar question on StackOverflow: Colorized grep -- viewing the entire file with highlighted matches

4
  • This doesn’t answer how to display all the input lines.
    – Chris Page
    Feb 15, 2020 at 3:17
  • @ChrisPage I would say it does: That's why each line gets printed but only "hello" is highlighted. Feb 16, 2020 at 7:07
  • To be pedantic — "the non-printable null string at the end of each line" — There is not a "non-printable null string" (0x00) at the end of lines, there is a line-terminator / end-of-line which will generally be 0x0d, 0x0a, or 0x0d0x0a (CR, LF, CRLF)
    – Stephen P
    Oct 1, 2020 at 0:28
  • 1
    You are confusing a null string with a null character (\0). Null string is an empty string. The term null string is taken from the regex documentation. See for example: man7.org/linux/man-pages/man7/regex.7.html Oct 1, 2020 at 21:04
8

Similarly to previous answer, you can catch all $ end of lines:

cat testfile | grep --color -E "hello|$"

-E (or --extended-regexp) means that special characters must be escaped with \. When using it, | will be treated as regex "OR" condition.

Grep |$ will also catch and print all lines which has an end, but since $ is a hidden character, it cannot be highlighted.

Update:

If you'd like to print all output, but also return exit code, whether match was found or not, you can use perl command:

cat testfile | \
perl -pe 'BEGIN {$status=1} END {exit $status} $status=0 if /hello/;'

If you prefer sed - Here's an example how to highlight all matches + return exit code if no match found: https://askubuntu.com/a/1200851/670392

4

Here's a solution using ripgrep, a modern, easier to use (IMO), and much faster replacement for grep:

cat my-file | rg --passthru '<REGEX>'

You can also easily customise the "highlighter colour":

# dark red foreground with light yellow background:
cat my-file | rg --passthru --colors match:fg:124 --colors match:bg:229 '<REGEX>'

See man rg (online version).


For everyday use I created a shell alias:

# in .bashrc
alias hl='rg --passthru --colors match:fg:124 --colors match:bg:229'

Now it's as simple as cat my-file | hl '<REGEX>'.

3

Adding to the top answer above. The highlight{} function mentioned in the comments only works when data is piped into it. The following alias, while not perfect, is more useful:

alias greph="grep --color=always -e^ -e"

This works with commands such as:

greph foo bar.txt

cat bar.txt | greph foo
0

-z is good to output the entire file as in the answer from bot779 but most times you need to control how many lines you want to include above and below the match. You can control the length of context with -C <N> where N is the number of lines of context. See the example below:

cat >testfile <<EOM
hello
there
my
friends
more
lines
to demonstrate
context
EOM

cat testfile | grep --color=always -C 2 friends

there
my
*friends* highlighted
more
lines

Using a large number such as -C 1000 can also limit the output if the output you are filtering is too big.

0

There is a much easier way to do this than using the --color option as all the other answers suggest. Use the tee command. Example:

cat my_file.txt | tee >(grep 'search_for_this')

This will display the full contents of my_file.txt and all instances of search_for_this will be highlighted.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .