Home / ubuntu / Questions / 916557
Accepted
dmx
Asked: 2017-05-19 10:49:26 +0800 CST

How can I use `awk` to split text in column?

  • 772

How can I extract first column (the first column in "")

"xxxxx1" {685a}
"xx2" {bcdd}
"xx3 gsdd" {29a6ff}
"sdsdf xxx" {243b9}
"sdfsdf ccc dd" {c28f2f}
"dsdsf sfsdf" {216e}
"sdfsdfsd" {48530}
"sdfsdff" {9d2afa0n}
"sdfsdfdff sdfs" {d8681a}
"sdfsdsds sdfsdf d" {5b9b8}
"sdfsdfs sdf sdfsdf" {68e08a}
"sdfsdfsdf sdf" {107fa0}

what I want to have as result is this: 

"xxxxx1"
"xx2"
"xx3 gsdd"
"sdsdf xxx"
"sdfsdf ccc dd"
"dsdsf sfsdf"
"sdfsdfsd"
"sdfsdff"
"sdfsdfdff sdfs"
"sdfsdsds sdfsdf d"
"sdfsdfs sdf sdfsdf"
"sdfsdfsdf sdf"

any idea?

command-line

4 Answers

  1. Best Answer

    Use ​ { (space, brace) as the field delimiter, since you don't care about the second field:

    $ awk -F ' {' '{print $1}' foo
"xxxxx1"
"xx2"
"xx3 gsdd"
"sdsdf xxx"
"sdfsdf ccc dd"
"dsdsf sfsdf"
"sdfsdfsd"
"sdfsdff"
"sdfsdfdff sdfs"
"sdfsdsds sdfsdf d"
"sdfsdfs sdf sdfsdf"
"sdfsdfsdf sdf"
    • 11

  2. If you just want everything between the first and last " double-quote character of each line, the most simple solution would probably be this, using grep instead of awk:

    grep -o '".*"' FILENAME

    The -o switch lets grep output only the matching parts instead of the whole line that contains the match. The (single-quoted, to prevent shell interpretation) pattern ".*" matches a sequence of any character (.) with any length (*), including zero, that is surrounded with double-quotes.

    • 3

  3. As an alternative to muru's awk solution.

    using cut:

    cut -f 1 -d { < file.txt

    using grep:

    grep -o '".*"' file.txt

    or

    grep -o \".*\" file.txt

    using sed:

    sed -r 's/(".*").*/\1/' file.txt

    or

    sed -r 's/\{.*\}//' file.txt

    or even:

    sed 's/{.*//'
    • 2

  4. Perl with grouping (.*) can do it too:

    $ perl -pe 's/"(.*)".*/"\1"/g' input.txt                                                    
"xxxxx1"
"xx2"
"xx3 gsdd"
"sdsdf xxx"
"sdfsdf ccc dd"
"dsdsf sfsdf"
"sdfsdfsd"
"sdfsdff"
"sdfsdfdff sdfs"
"sdfsdsds sdfsdf d"
"sdfsdfs sdf sdfsdf"
"sdfsdfsdf sdf"

    The trick here is that we match whole line, and use "(.*)" to treat everything between double quotes as one group. We replace that whole line with the group we matched by referring to it via \1 part.

    • 1