read values of txt file from bash

Question

I'm trying to read values from a text file.

I have test1.txt which looks like

sub1    1   2   3
sub8    4   5   6

I want to obtain values '1 2 3' when I specify 'sub1'.

The closest I get is:

subj="sub1"
grep "$subj" test1.txt

But the answer is:

sub8    4   5   6

I've read that grep gives you the next line to the match, so I've tried to change the text file to the following:

test2.txt looks like:

sub1    
1   2   3

sub8    
4   5   6

However, when I type

grep "$subj" test2.txt

The answer is:

sub1

It should be something super simple but I've tried awk, seg, grep,egrep, cat and none is working...I've also read some posts somehow related but none was really helpful

Answer 1

Awk works: awk '$1 == "'"$subj"'" { print $2, $3, $4 }' test1.txt

The command outputs fields two, three, and four for all lines in test1.txt where the first field is $subj (i.e.: the contents of the variable named subj).

Answer 2

grep "sub1" test1.txt | cut -c6-

or

grep -A 1 "sub1" test2.txt | tail -n 1

Answer 3

With your original text file format:

target=sub1
while IFS=$' \t\n' read -r key values; do
  if [[ $key = "$target" ]]; then
    echo "Found values: $values"
  fi
done <test1.txt

This requires no external tools, using only functionality built into bash itself. See BashFAQ #1.

---

As has come up during debugging in comments, if you have a traditional Apple-format text file (CR newlines only), then you might want something more like:

target=sub1
while IFS=$' \t\n' read -r -d $'\r' key values || [[ $key ]]; do
  if [[ $key = "$target" ]]; then
    echo "Found values: $values"
  fi
done <test1.txt

---

Alternately, using awk (for a standard UNIX text file):

target="sub1"
awk -v target="$target" '$1 == target { $1 = ""; print; }' <test1.txt

...or, for a file with CR-only newlines:

target="sub1"
tr '\r' '\n' <test1.txt | awk -v target="$target" '$1 == target { $1 = ""; print; }'

This version will be slower if the text file being read is small (since awk, like any other external tool, takes time to start up); but faster if it's large (since awk's operation is much faster than that of bash's built-ins once it's done starting up).

Answer 4

Sed also works: sed -n -e 's/^'"$subj"' *//p' file1.txt

It outputs all lines matching $subj at the beginning of a line after having removed the matching word and the spaces following. If TABs are used the spaces should be replaced by something like [[:space:]].

Answer 5

There's a bunch of ways to do this (and shorter/more efficient answers than what I'm giving you), but I'm assuming you're a beginner at bash, and therefore I'll give you something that's easy to understand:

egrep "^$subj\>" file.txt | sed "s/^\S*\>\s*//"

or

egrep "^$subj\>" file.txt | sed "s/^[^[:blank:]]*\>[[:blank:]]*//"

The first part, egrep, will search for you subject at the beginning of the line in file.txt (that's what the ^ symbol does in the grep string). It also is looking for a whole word (the \> is looking for an end of word boundary -- that way sub1 doesn't match sub12 in the file.) Notice you have to use egrep to get the \>, as grep by default doesn't recognize that escape sequence. Once done finding the lines, egrep then passes it's output to sed, which will strip the first word and trailing whitespace off of each line. Again, the ^ symbol in the sed command, specifies it should only match at the beginning of the line. The \S* tells it to read as many non-whitespace characters as it can. Then the \s* tells sed to gobble up as many whitespace as it can. sed then replaces everything it matched with nothing, leaving the other stuff behind.

BTW, there's a help page in Stack overflow that tells you how to format your questions (I'm guessing that was the reason you got a downvote).

-------------- EDIT ---------

As pointed out, if you are on a Mac or something like that you have to use [:alnum:] instead of \S, and [:blank:] instead of \s in your sed expression (as these are portable to all platforms)

Answer 6

You doing it right, but it seems like test1.txt has a wrong value in it.

with grep foo you get all lines with foo in it. use grep -m1 foo to find the first line with foo in it only.

then you can use cut -d" " -f2- to get all the values behind foo, while seperated by empty spaces.

In the end the command would look like this ...

$ subj="sub1"
$ grep -m1 "$subj" test1.txt | cut -d" " -f2-

But this doenst explain why you could not find sub1 in the first place. Did you read the proper file ?

Answer 7

awk '/sub1/{ print $2,$3,$4 }' file
1 2 3

What happens? After regexp /sub1/ the three following fields are printed. Any drawbacks? It affects the space.