0

I have file abc.txt which has data like -

abc pqr testing, xyz=a432j42jjk4, key=124

abc pqr testing, xyz=jkl234j54nn, key=567
abc pqr testing, xyz=2395hdshkw4, key=3232


abc pqr testing, xyz=abc424729hh, key=7676
abc pqr testing, xyz=70700ghgh99, key=12342

I am searching text "xyz" using awk like -

awk -F"xyz=" '{print $2}' abc.txt | awk '{print $1}'

which gives me output with new lines in it -

a432j42jjk4,

jkl234j54nn,
2395hdshkw4,


abc424729hh,
70700ghgh99,

I am looking for shortcut or trick in awk where I can get output at the same time remove newlines from it.

Improve this question
1
  • The trick that you're looking for is to put NF as the pattern to match for either rule. For example, use NF {print $2} or NF {print $1} or both in your case. Placing NF makes sure that only non-blank lines/records are processed by the rule. Commented Jan 25, 2019 at 8:12

3 Answers 3

Reset to default
1

Use GNU grep for it:

$ grep -o -P '(?<=\bxyz=)[^,]+,' abc.txt

grep prints lines that matches a given regular expression. With -o we define to print only the matching part of the line. With -P our regex is interpreted as a perl regex, which give us more option.

Our regex says:

  • find one more characters that are not a , ([^,]+) and end with ,
  • these consecutive character should follow after a xyz= (?<=\bxyz=) this is called a positive lookbehind). With the \b we make sure not grep something like booxyz=
Improve this answer
7
  • how this exactly works. can you please elaborate significance of each options & regex? Commented Jan 25, 2019 at 5:53
  • also this prints ) [^ along with the output. how I can ignore that? Commented Jan 25, 2019 at 6:17
  • 1
    a positive lookbehind to be more precise Commented Jan 25, 2019 at 6:25
  • 1
    grep doesn't add any additional character. So I don't understand what you mean by "this prints ) [^ along with the output"? Commented Jan 25, 2019 at 6:32
  • Output was correct but first two lines were ) [^. I added sed '/^) /d' and removed Commented Jan 25, 2019 at 7:04
1

We can use the built-in match capability of awk to streamline the command further:

$ awk -F ',' '/xyz=/ {split($2,a,"="); print a[2]}' file
a432j42jjk4
jkl234j54nn
2395hdshkw4
abc424729hh
70700ghgh99

Here we use awk to match lines that contain xyz=, so we don't have to use the if statement later to strip out the empty lines.

This looks more readable because the logic flows more from left to right.

This solution does not guard against the case where xyz= appears in a field other than the second in the comma-delimited file.

Improve this answer
0

Using standard awk and assuming that you'd want to look for xyz= specifically in the second comma-delimited field.

$ awk -F ',' '{ split($2,a,"="); if (a[1] == " xyz") print a[2] }' file
a432j42jjk4
jkl234j54nn
2395hdshkw4
abc424729hh
70700ghgh99

Alternatively,

$ awk -F ',' '{ split($2,a,"=") } a[1] == " xyz" { print a[2] }' file
a432j42jjk4
jkl234j54nn
2395hdshkw4
abc424729hh
70700ghgh99

This would treat the input as consisting of lines with comma-delimited fields. The actual body of the awk code will split the second comma-delimited field into subfields on =, and then print the second one of these if the first one is exactly the string xyz (including the initial space before the comma).

Empty lines would be dropped as a[1] would not get the wanted value on these lines.

Would you want extra commas at the end of each line of output, use print a[2] ",".

Improve this answer

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.