CODEWITHSUNDEEP
regexawksedgrep
ogbofjnr
ogbofjnr

Reputation: 2028

How to extract all variables from file

I have a file with the following format:

text${v1}text${v2}etc

How do I extract all variables names from it using bash?

Expected output is: v1 v2

Upvotes: 3

Views: 267

Answers (5)

RARE Kpop Manifesto
RARE Kpop Manifesto

Reputation: 2865

echo 'text${v1}text${v2}etc' | 

mawk 'gsub(OFS"+", __, $!(NF = NF)) + \
      gsub("^"(__) "|"(__)"$", _)^_ + gsub(__," /\f\b\b/ ")' \
                                            __='\300'                                    \
  FS='(^[^$}{]*)?[$][{]|[}][^}{$]*|[\301]+' OFS='\301'

v1 /
  / v2

Upvotes: 1

The fourth bird
The fourth bird

Reputation: 163477

An alternative with grep and -P to enable Perl-compatible regular expressions:

grep -oP '\${\K[^}]*(?=})' file

The pattern matches:

  • \$ Match $
  • { Match literally
  • \K Forget what is matched until now
  • [^}]* Match optional repetitions of any char except }
  • (?=}) Positive lookahead, assert } to the right

Output

v1
v2

Or using gnu awk with a capture group in the regex ([^}]*) and print that group 1 value using a[1] (same looping mechanism as the answer of @RavinderSingh13)

awk '
{
  while(match($0,/\${([^}]*)}/, a)){
    print a[1]
    $0=substr($0,RSTART+RLENGTH)
  }
}
' file

Upvotes: 4

sseLtaH
sseLtaH

Reputation: 11237

Using sed

sed -E 's/[^$]*\$\{([^}]*)}|[[:alpha:] ]+$/\1 /g' input_file
v1 v2

Upvotes: 4

RavinderSingh13
RavinderSingh13

Reputation: 133650

1st solution: Within single GNU awk you could try following code. Written and tested with shown samples only. Simple explanation would be, setting RS(record separator) as \\${[^}]*} and then in main program substituting values like ${ and } from RT and printing only required values as per requirement.

awk -v RS='\\${[^}]*}' 'RT{gsub(/^\${|}$/,"",RT);print RT}'  Input_file

2nd solution: With any awk please try following awk code.

awk '
{
  while(match($0,/\${[^}]*}/)){
    print substr($0,RSTART+2,RLENGTH-3)
    $0=substr($0,RSTART+RLENGTH)
  }
}
'  Input_file

Upvotes: 6

anubhava
anubhava

Reputation: 785641

Using simple grep + sed:

grep -o '\${[^}]*}' file | sed -E 's/^\${|}$//g'

v1
v2


Here:

- `grep` matches and prints `${...}` strings
- `sed` removes `\${` and `}` from output

Upvotes: 5

Related Questions