Clever Geek Handbook

Awk - delete a line if the following lines do not match the pattern

I would like to solve the following problem and so far I think awk might be right too, but please confirm if I am wrong.

I have console output that follows this pattern:

HEADER: HeaderInfo1
Details 1
Details 2
HEADER: HeaderInfo2
HEADER: HeaderInfo3
Details 3
Details 4
HEADER: HeaderInfo4

I would like to convert this to the following:

HeaderInfo1: Details 1
HeaderInfo1: Details 2
HeaderInfo3: Details 3
HeaderInfo3: Details 4

So, as you can see, every detail belongs to the nearest title. Number of details in a variable header Empty headers should be ignored.

I've tried all sorts of regex and sed magic, but this is similar to the awk case I guess. However, I don't quite know where to start.

+3


source to share


1 answer


Assuming your information is in a file named info

:

$ awk '/^HEADER/{hdr=$2;next} {print hdr": "$0}' info
HeaderInfo1: Details 1
HeaderInfo1: Details 2
HeaderInfo3: Details 3
HeaderInfo3: Details 4

Explanation



The program awk

is divided into two parts:

  • /^HEADER/{hdr=$2;next}

    For lines starting with HEADER

    , the second field is stored in a variable hdr

    and then awk

    instructed to move to the next line.

  • {print hdr": "$0}

    For all other lines, the last header value is printed hdr

    , followed by a colon, followed by a line ( $0

    ).

+4


source






More articles:


All Articles