Finding exact string in a file using a variable (grep)

Question

I am writing a script currently that is searching inside a txt file.

So basically I have a loop that grabs a name from a txt file and then searches for that name in another file (file2), but I am having issues.

name=flex

grep $name file2 > file2output

So if file2 (file I am searching inside of) has something like

    1 flex-inside

    2 flex

    1 flex-end

It will match all three of those when I only want it to match the exact string flex.

I want it to only match

2 flex

I was trying to do something where it would look for whatever is in $name and then \n. But nothing is working.

Thank you for the help.

Now try grep -Fw "$name" file. And, don't just say "no luck." That's useless. Tell us exactly what happened: Was there an error? If so, show us the complete error message. Was there too much or too little output? If so, what exactly what was the output? — John1024
– John1024, Commented Feb 15, 2018 at 21:58
If grep -Fw "$name" produces output but grep -Fx "$name" file does not, that means that you have characters, invisible or otherwise, before or after flex in that line. If grep "$name$" file works for you, that means that those characters, visible or not, are before flex. Try running hexdump -C file and see what extra characters are on the flex line. — John1024
– John1024, Commented Feb 15, 2018 at 22:24
It sounds like the word you are looking for will always appear at the end of the line. In that case, try grep " $name$" file where there is a space before $name and, the final $ is the regex symbol for end-of-the-line. — John1024
– John1024, Commented Feb 15, 2018 at 23:50
No, that won't work. You wanted only whole word matches. Without the space in the regex, it will match aflex or bflex as well flex — John1024
– John1024, Commented Feb 16, 2018 at 0:03

John1024 · Accepted Answer · 2018-02-16 03:58:37Z

1

As per the comments, the input file looks like:

$ cat file2
    1 flex-inside
    3 reflex
    2 flex
    1 flex-end

We want to select the line matching flex but not reflex or flex-end, etc. In that case, try:

$ name='flex'
$ grep " $name$" file2
    2 flex

There are two key points here:

There is a space before $name to assure that words like reflex are not matched.
There is a dollar sign, $, after $name to require that flex occurs at the end of the line.

Extended version: handling embedded spaces

The above all assumes that the matched text after the numbers does not include spaces. Consider for example the input file:

$ cat file2
    1 flex inside
    3 re flex
    2 flex
    1 flexible

We can select the one line matching flex with:

$ grep -E "^[[:space:]]*[[:digit:]]+[[:space:]]+$name$" file2
    2 flex

Here, ^[[:space:]]*[[:digit:]]+[[:space:]]+ matches zero or more spaces that begin the line followed by digits followed by one or more spaces. This is followed by $name$ which matches $name appearing at the end ($) of the line.

The use of character classes like [:space:] and [:digit:] assures that this code is unicode safe.

answered Feb 16, 2018 at 3:58

John1024

115k15 gold badges152 silver badges183 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Tursko Over a year ago

Thank you so much. Great explanation.

Collectives™ on Stack Overflow

Finding exact string in a file using a variable (grep)

1 Answer 1

Extended version: handling embedded spaces

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Extended version: handling embedded spaces

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related