-n --> provides you a while loop by reading records (lines (terminated by new-lines) by default)
from the filename arguments without default printing (similar to sed -n).
-e --> signifies that the next argument is a script/program.
-l --> does automatic line-ending processing. By itself (and with default input separator as
new-line), will cause print to automatically add a new-line (spcifically the output
record separator which has been assigned the value of current input record separator
during this switch's processing (Phew...!!!)) after its arguments.
I'll just explain the first one. Hope you can make out the second one on your own.
Read each line of input and try to match the pattern recursively.
I'll rewrite the snippet with the /x modifier to the match operator for explanation:
Code:
perl -ne 'while(
/
( ### capturing group 1 starts
(HOST|PORT) ### capturing group 2 starts and ends - matches either HOST or PORT
[ ]=[ ] ### a space followed by = and then a space
### had to use bracket expressions because white spaces are ignored with x
### modifier
.+? ### 1 or more occurrences of any character (.+), non-greedy matching (?)
### non-greedy matching starts matching from the least amount of characters
### possible, as opposed to greedy matching. This ensures that you do not match
### unintended strings.
) ### capturing group 1 ends
\) ### a literal closing parenthesis
/xg
) {print "$1\n"}' tnsnames.ora
This also exploits a peculiarity/feature of the match operator with the /g modifier.
The match operator with the /g modifier keeps track of where it has matched a pattern in a string
(rather it is the string which keeps track of the same, check perldoc -f pos for details). So, if you
try to match again, it starts from the correct offset.
for the first iteration, capturing group 1 matches HOST = atoadb01 and you print it out with $1.
for the second iteration, matching starts from where the 1st one left off. So, capturing group 1 now matches PORT = 1521.
for the third iteration, no match is possible after the second match and the match operator returns false
in the scalar context and you stop processing that line and proceed with the next one.
Last edited by elixir_sinari; 02-14-2013 at 05:50 AM..
This User Gave Thanks to elixir_sinari For This Post:
(1) The parentheses "()" without escape back-slashes capture the group of characters and create a backreference.
(2) A parenthesis with an escape back-slash, i.e. either "\(" or "\)" matches a literal parenthesis.
(3) .* matches the maximum sequence of characters (greedy search).
(4) .*? matches the minimum sequence of characters (non-greedy search).
Here's how the regex works. The color of a portion of the regex is the same as the text it matches.
(1)
Code:
/(HOST.*)\)\((.*?)\)/
Capture the maximum sequence of characters starting with "HOST".
The following could potentially be stored in $1 so far.
Attempting to contact (DESCRIPTION = (ADDRESS = (PROTOCOL = TCP)(HOST = atoadb01)(PORT = 1521)) (CONNECT_DATA = (SERVER = DEDICATED) (SERVICE_NAME = atta1)))
(2)
Code:
/(HOST.*)\)\((.*?)\)/
Capture the maximum sequence of characters starting with "HOST",
- followed by the literal character ")"
The following could potentially be stored in $1 so far.
Attempting to contact (DESCRIPTION = (ADDRESS = (PROTOCOL = TCP)(HOST = atoadb01)(PORT = 1521)) (CONNECT_DATA = (SERVER = DEDICATED) (SERVICE_NAME = atta1)))
(3)
Code:
/(HOST.*)\)\((.*?)\)/
Capture the maximum sequence of characters starting with "HOST",
- followed by the literal character ")"
- followed by the literal character "("
The following could potentially be stored in $1 so far.
Attempting to contact (DESCRIPTION = (ADDRESS = (PROTOCOL = TCP)(HOST = atoadb01)(PORT = 1521)) (CONNECT_DATA = (SERVER = DEDICATED) (SERVICE_NAME = atta1)))
(4)
Code:
/(HOST.*)\)\((.*?)\)/
Capture the maximum sequence of characters starting with "HOST",
- followed by the literal character ")"
- followed by the literal character "("
- followed by the minimum sequence of characters till the literal character ")".
The following could potentially be stored in $1 and $2 so far.
Attempting to contact (DESCRIPTION = (ADDRESS = (PROTOCOL = TCP)(HOST = atoadb01)(PORT = 1521)) (CONNECT_DATA = (SERVER = DEDICATED) (SERVICE_NAME = atta1)))
Since the pattern has been parsed, we have the values of $1 and $2 now, which we could print.
tyler_durden
This User Gave Thanks to durden_tyler For This Post:
Thanks a tonne elixir_sinari and durden_tyler for the effort and the time that you spent on explaining this code !! Your lucid explanations are very intuitive and give a clear picture.
@durden_tyler, I just have one trivial question. Why in the script the non-greedy regular expression character '?' is ignored for the "HOST" pattern matching string (when I include it, I get the same correct output yet eager to know why it is not included) also what wrong in explicitly mentioning the "PORT" pattern-matching string in the 2nd back-reference group?
Code:
perl -lne 'print "$1\n$2" if /(HOST.*?)\)\((PORT.*?)\)/' tnsnames.ora
Last edited by royalibrahim; 02-15-2013 at 05:39 AM..
#!/bin/bash
n=$l; typeset -a v
x=$(< input.dat)
check(){
if; then
sed 's/Test/Proc/g' file.sh >fl.sh
else
exit 13
fi
}
check $n
while ; do
x=`expr $x -l`
v=$x
done
less fi.sh l>/dev/null&& echo yes || exit 1
echo v= ${v
}
exit 0
I have file.sh and input.dat in the current... (3 Replies)
Hi,
I am breaking my head to understand the below line of code in perl:
next if $pid = fork; # Parent goes to next server.
die "fork failed: $!" unless defined $pid;
Can anyone please explain me in detail as I am a slow learner?
Thanks,
Ali. (2 Replies)
Hi All,
Good Morning !
Can anyone help me understand the following scenario.
Scenario 1 : 2 LINUX Scripts(Process 1 & 2) are trying to copy the same file (lets call it FileA.txt) to the same directory (lets say, /home/abc/zyz) at the same time. If both copy is happening at the same... (9 Replies)
Hi,
I have a requirement as follows. Have 3 files. Need to match up the data in each one of them and sum up the data by a field and display it. example given below.
File 1 : Name, Emp id
File 2 : Empid, Subject,
File 3 : Subject, Score, Class
Match Emp id in File 1 and File 2 and then... (7 Replies)
I am debugging a script and have stuck up at one code line
awk -F , '{if (match($3,001)) { print $2 } }' Master20120307090511.tmp
The Master20120307090511.tmp is
001,ARE , 001
002,ARE , 002
003,ARE , 003
006,ARE , 006
011,ARE , 011
012,ARE , 012
What happens is when i fire this ... (5 Replies)
Hi,
I am hitting my head for weeks now,
I run 3 sites that are basically serving 1 community, a job search, a wiki and a social networking platform.
Rather than promote all three I want to create the illusion that its all 1 website, and I identify 3 areas to make this possible.
1) unified... (7 Replies)
Hello,
In my system, I have implemented various alerts.
An example:
#!/bin/sh
BILL=bill@company.com
#GETS THE CURRENT NO OF SESSIONS OPENED IN ORACLE
SESSION_NO=`sqlplus -s ${TEST_USER} <<EOF
SET HEAD OFF
@/home/alerts/vlib/cur_session_no.sql
EOF`
#MANUALLY SET A THRESHOLD FOR THE... (1 Reply)
hi guys linux noob here wanting to learn linux scripting, i need help with a backup script that not only allows me to back up my files but restore them to my own personal directory, had a look at some of the coding from the scripts section on this site but still lost.
any help is much appreciated (8 Replies)
Hi,
I have doubt regarding the Anonymous hash. For creating a object we are generally using Anonymous hash.
My Doubt is:
Why we are only using Anonymous hash?? Instead of Anonymous hash can we use global hash variable and take its reference for creating an object using same bless function??... (0 Replies)
Hello All,
Currently i am working on a logic in PERL scripting ...
But i am sort of stuck up, can any one please help.
Here goes.
1. Search for a pattern in a file
2. If the pattern matched lets say 10 lines
2.1 Reterive the first line and check for another pattern
2.1.1 if... (1 Reply)
02-14-2013
!! Your lucid explanations are very intuitive and give a clear picture.