CODEWITHSUNDEEP
parsingunixawk
Suri Bellen
Suri Bellen

Reputation: 21

Parse a file to pull out 2 sets of information

I have a log file that keeps a record of user inputs. Each line in the log is unique and I need to pull out 2 specific items - a userId and a URL. I can't just use awk < file '{print$1, print$6}' because the items are not always in the same position in each line.

Example text:

userId='1' managed:no address:123street phone:1234567890 http:/someurl.com
newuser:yes userId='2' managed:yes address:123street  http:/someurl.com
userId='3' address:123 street phone:1234567890 http:/someurl.com
userId='4' managed:no address:123street phone:1234567890 http:/someurl.com

I need to parse the userId and URL address to a file, but these are not always in the same position in each line. Any suggestions would be greatly appreciated.

Upvotes: 2

Views: 52

Answers (3)

Ed Morton
Ed Morton

Reputation: 203229

$ awk '{for(i=1;$i!~/userId/;i++); print $i, $NF}' file
userId='1' http:/someurl.com
userId='2' http:/someurl.com
userId='3' http:/someurl.com
userId='4' http:/someurl.com

Upvotes: 2

Gilles Qu&#233;not
Gilles Qu&#233;not

Reputation: 185015

Try the following code :

gawk '{
    for (i=1; i<=NF; i++)
        if ($i ~ "^userId=") id=gensub(/userId=\047([0-9]+)\047/, "\\1", "", $i)
        else if ($i ~ "^http") url=$i
        print "In line "NR", the id is "id" and the url is "url
}' file.txt

Sample input :

userId='1' managed:no address:123street phone:1234567890 http:/someurl1.com
newuser:yes userId='2' managed:yes address:123street  http:/someurl2.com
userId='3' address:123 street phone:1234567890 http:/someurl3.com
userId='4' managed:no address:123street phone:1234567890 http:/someurl4.com

Sample output :

In line 1, the id is 1 and the url is http:/someurl1.com
In line 2, the id is 2 and the url is http:/someurl2.com
In line 3, the id is 3 and the url is http:/someurl3.com
In line 4, the id is 4 and the url is http:/someurl4.com

This solution have the advantage to have id or http items to be anywhere you'd want in lines.

Upvotes: 1

hek2mgl
hek2mgl

Reputation: 157947

With awk:

awk '{for(c=1;c<NF;c++){if(match($c,/userId/)){print $c,$NF; break}}}' your.file

Output:

userId='1' http:/someurl.com
userId='2' http:/someurl.com
userId='3' http:/someurl.com
userId='4' http:/someurl.com

Upvotes: 0

Related Questions