joining each line of the file with underscore

Question

I have an input file named test.txt like this:

Homo sapiens
Mus musculus
Rat rattus rattus

I want to use a for loop here and loop through each line. Some thing like this:

for i in "`cat test.txt`"
do
        fn=store_name
        do something else > ${fn}.txt
done

Since each line has a space between names I used double quotes in cat test.txt. I want to make a variable fn which attaches _ between these names of the lines so my fn becomes Homo_sapiens and Mus_musculus and if I have Rat rattus rates it becomes Rat_rattus_rattus

How can I achieve this?

I have to 2nd that comment, don't use for to read lines, it's just over complicating it when there are so many cleaner safer ways to do this. — slm, Aug 01 '18 at 20:27

score 5 · Answer 1 · answered Aug 01 '18 at 19:12

5

Use tr to transliterate a space to an underscore:

while read line
do
    echo ${line} | tr -s " " "_"
done < test.txt

Homo_sapiens Mus_musculus Rat_rattus_rattus

answered Aug 01 '18 at 19:12

JRFerguson

14,740

jesse_b · Answer 2 · 2018-08-01T19:11:35.300

4

Using awk:

awk '{OFS="_";}{$1=$1; print}' test.txt

Using a loop:

while read -r line; do 
  line=${line// /_} 
  command using "$line"
done<test.txt

edited Aug 01 '18 at 19:11

answered Aug 01 '18 at 19:09

jesse_b

37,005

this works but I have to operate also on those underscore names within the loop – user3138373 Aug 01 '18 at 19:10
@user3138373: I've updated it to include a loop solution – jesse_b Aug 01 '18 at 19:12
@user3138373 - why would you need a "loop" ? Text should never be processed via "loops". – don_crissti Aug 01 '18 at 22:21

glenn jackman · Answer 3 · 2018-08-01T20:54:02.597

2

I'd use sed and read its output with a loop:

sed 's/[[:space:]]\+/_/g' test.txt | while IFS= read -r line; do 
    process "$line" however
done > some.output.file

edited Aug 01 '18 at 20:54

answered Aug 01 '18 at 20:25

glenn jackman

85,964

score 0 · Answer 4 · answered Aug 01 '18 at 20:27

0

Here's a quick for loop that uses sed to get it done:

for file in $(sed "s/ /_/g" test.txt); do echo "test" > $file; done

Just replace the echo command with whatever it is you want to do.

I use commands like this all the time though out the day.

answered Aug 01 '18 at 20:27

Jeff Aigner

1

Beam Davis · Answer 5 · 2018-08-01T22:16:41.890

-1

Someone gave an answer with tr and someone else gave one with awk, now here's how to do it with sed:

for i in "`cat test.txt`"
do
    fn="$(echo "$i" |sed -e 's/ /_/g')"
    'do something else' > $fn.txt
done

edited Aug 01 '18 at 22:16

answered Aug 01 '18 at 19:34

Beam Davis

199

My point being there are about a hundred ways to do this...and he did tag his question with "sed". – Beam Davis Aug 01 '18 at 19:44
1

I don't understand why you are using a for loop instead of while read. Also why are you modifying IFS instead of just quoting your command substitution like op has done? Finally, you should use $( ... ) over backticks. – jesse_b Aug 01 '18 at 19:54
1

don't read lines with "for" – glenn jackman Aug 01 '18 at 20:23
Modified to use quotes instead of IFS. As for "for" vs. "while", six one way, half-a-dozen the other. Also, nothing wrong with backticks if you have no quotes inside the command being called. – Beam Davis Aug 01 '18 at 22:14
Using "for" loop instead of "while" loop also doesn't matter in this case, because we don't want to handle blank lines anyway. – Beam Davis Aug 01 '18 at 22:22

joining each line of the file with underscore

5 Answers5