Extract and print string between two patterns

Question

The below is the content in a file

SM Filesystems - ci240min;vlg00457.wdf.sap.corp;00205495930028250313;;Virtual_Gcp,Virtual,SISM_responsible,Linux_Suse,Linux,E2E_Services,Database,DLM;;;;0;warning;OK: all local filesystems are writeable OK: /var is mounted with correct options \nOK: /home is mounted with correct options \nOK: /opt/bmc mounted and writeable \nOK: /boot disk usage 18% of 250M with available size of 195 MB\nOK: / disk usage 69% of 7.9G with available size of 2522 MB\nOK: /var disk usage 31% of 2.0G with available size of 1406 MB\nOK: /opt disk usage 43% of 4.0G \nOK: /tmp disk usage 4% of 2.0G with available size of 1945 MB\nOK: Found no max. files in any monitored directory \nOK: /boot/ inode usage 1% \nOK: / inode usage 45% \nOK: /var/ inode usage 4% \nOK: /opt/ inode usage 4% \nOK: /tmp/ inode usage 1% \nOK: Initialised Physical disks are in use \nERROR: Disks partitions sde to be checked \n

I need to print vlg00457.wdf.sap.corp and sde which is in between the strings "partitions" and "to" using single awk. Right now, I was able to get with two awk separated

cat bala.txt | grep -i "ERROR: Disks partitions" | tail -1 | awk -F ";" '{print $2}'
vlg00457.wdf.sap.corp


cat bala.txt | grep -i "ERROR: Disks partitions" | tail -1 | awk -F"partitions" '/partitions/{print $2}'
 sde to be checked \n

echo "Disks partitions sde to be checked" | awk '{print $3}'? — Cyrus, Oct 26 '19 at 13:58
@BalajiNavaneedhakrishnan: Please edit your question with the details you have provided in comments and we can reopen/answer it. — jesse_b, Oct 26 '19 at 16:54
awk -F';' '{ printf "%s ", $2; sub(/^.*ERROR: Disks partitions /,""); sub(/ /,";"); print $1}'. or to do it entirely in awk (without cat, grep, or tail): awk -F';' '/ERROR: Disks partitions/ { a=$2; sub(/^.*ERROR: Disks partitions /,""); sub(/ /,";"); b=$1} END { if ( a != "" && b != "") {print a,b}}' bala.txt — cas, Oct 27 '19 at 23:54
This is unclear. (1) The input file is unreadable. Is it really one line that’s 888 characters long, or does \n represent newlines? (2) How is “vlg00457.wdf.sap.corp” matched? (3) Are string1 and string2 guaranteed to both be present on the same line, in the correct order? (4) Can there be multiple string1 / string2 pairs in the file? (5) Can string1, string2, string3 and string4 appear on the same line? (6) Is the text to be extracted and printed (between the strings) guaranteed to be exactly one word? — G-Man Says 'Reinstate Monica', Oct 28 '19 at 18:13
Possible duplicate of extract value between two search patterns on same line — G-Man Says 'Reinstate Monica', Oct 28 '19 at 18:13
See also How to find / grep what is between string1 and string2?  and  How do I display all the characters between two specific strings? — G-Man Says 'Reinstate Monica', Oct 28 '19 at 18:13

Jim L. · Answer 1 · 2019-10-30T16:10:24.223

1

Using sed:

$ grep "ERROR: Disks partitions" bala.txt |
    sed -E 's/(^.*ERROR: Disks partitions | to be checked.*$)//g'
sde

edited Oct 30 '19 at 16:10

answered Oct 28 '19 at 19:35

Jim L.

7,997
1
13
27

This will work (for the “sde” part) if the file, as shown, is a single line (with 888 characters). But those characters contain 16 occurrences of “\n”. If those are meant to represent newlines, then it is a 16-line file, and your answer will print the first 15 lines in their entirety. – G-Man Says 'Reinstate Monica' Oct 29 '19 at 02:10
Thank you, @G-Man. I have improved my answer based on your feedback. – Jim L. Oct 30 '19 at 16:11

score 0 · Answer 2 · answered Oct 28 '19 at 16:54

0

piece of cake with Perl:

$ perl -F\; -lane '/partitions (\S+)/ and print "$F[1]\t$1"' bala.txt

answered Oct 28 '19 at 16:54

Rakesh Sharma

836

Extract and print string between two patterns

2 Answers2