Bash: how to get lines between patterns only if there is pattern2 between them

Question

The situation is next. I got log file where logs separated with minus chars like:

Timestamp1
---
Log: 1
Address: http://addr1.com
Payload: <soap:Envelope>
             <soap:Body>
                 <context 1-1>
                 <context 1-2>
                 <context 1-3>
             </soap:Body>
         <soap:Envelope>
---;
Timestamp2
---
Log: 2
Address: http://addr2.com
Payload: <soap:Envelope>
             <soap:Body>
                 <context 2-1>
             </soap:Body>
         <soap:Envelope>
---;
Timestamp3
---
Log: 3
Address: http://addr3.com
Payload: <soap:Envelope>
             <soap:Body>
                 <context 3-1>
                 <context 3-2>
             </soap:Body>
         <soap:Envelope>
---;
...

I need to get whole log info where some keyword found e.g. if keyword is "context 2-1" next strings should be printed:

---
Log: 2
Address: http://addr2.com
Payload: <soap:Envelope>
             <soap:Body>
                 <context 2-1>
             </soap:Body>
         <soap:Envelope>
---;

So how can i do this pattern search with "greedy" cut of delimeters around it?

Remove `...` and show better sample and also show your attempts. — anubhava, Oct 10 '16 at 14:02

SLePort · Accepted Answer · 2016-10-10T17:11:23.800

2

With sed:

sed -n '/^---/ {:a;N;/---;/!ba;/context 2-1/p}' file

Explanation:

/^---/ when line starting with --- is found
a: label for to come loop
N: add next line to the pattern space
/---;/!: while ---; if not found...
ba loops to a label
/context 2-1/p: when loop terminates, print all lines previously added to pattern space if context 2-1 is found

edited Oct 10 '16 at 17:11

answered Oct 10 '16 at 16:17

SLePort

15,211
3
34
44

score 0 · Answer 2 · edited May 23 '17 at 10:34

Using this as a guideline: How to select lines between two patterns?

$0=="---" {                   # at the fron marker
    flag=1                    # flag up
    buf=""                    # clear the buf
} 
$0=="---;" { flag=0 }         # at the end marker, flag down
{
    buf=buf $0 (flag?RS:"")   # gather buffer, add RS before the end marker
    if($0 ~ "^Payload2")      # if keyword found
        output=1              # output flag up
} 
flag==0 && output==1 {        # after end marker when putput flag up
    print buf                 # output buf
    output=0                  # output flag down
}

Run it:

$ awk -f script.awk logfile
---
Log2
Address2 ...
Payload2 ...
---;

Thank you for the answer. Yes, i know how to cut text between patterns using sed, but i didn't know how to check another existence of second pattern. — Vadim K, Oct 10 '16 at 15:08

score 0 · Answer 3 · edited Oct 10 '16 at 23:40

0

awk -vRS="Timestamp[0-9]+"  -v k="context 2-1"  '$0~k' file2

This uses Timestamp[0-9]+ as a new-line character. k is keyword what you want. If $0 matches the keyword, then print $0.

edited Oct 10 '16 at 23:40

Scott Weldon

9,673
6
48
67

answered Oct 10 '16 at 15:05

zxy

148
1
2

Bash: how to get lines between patterns only if there is pattern2 between them

3 Answers3