I have a psv file having records something like this:

  Apple    | Banana     | I want to eat banana        | Carrot

As you can see the all the columns contain leading and trailing spaces I used this command:

sed -i -r 's/\s //g;/^$/d' fileName

to remove the blank spaces as well as blank lines. After which i got the result as:

Apple|Banana|Iwanttoeatbanana|Carrot

What my expected Result is somewhat like this:

Apple|Banana|I want to eat banana|Carrot

Please help me getting this desired result using shell script.

Thank you in advance!

CodePudding user response：

You may use this sed:

sed -E 's/^[[:blank:]] |[[:blank:]]*(\|)[[:blank:]]*|[[:blank:]] $/\1/g' file

Apple|Banana|I want to eat banana|Carrot

Explanation:

• ^[[:blank:]] : Match 1 or more whitespaces after start
• |: OR
• [[:blank:]]*(\|)[[:blank:]]*: Match 0 or more whitespaces before and after a | and capture | in group #1
• |: OR
• [[:blank:]] $: Match 1 or more whitespaces before end
• Replacement is a \1` that puts captured value of group #1 in replacement

An awk solution:

awk -F '[[:blank:]]*\\|[[:blank:]]*' -v OFS='|' '{
   gsub(/^[[:blank:]] |[[:blank:]] $/, ""); $1=$1} 1' file

Apple|Banana|I want to eat banana|Carrot