topic grep help in Operating System - HP-UX

grep help

George Chechakunnil — Tue, 28 Jul 2009 09:39:46 GMT

Hi

i have a file with series of entries with 4 lines each in a group

How can i search for a pattern and print all the 4 lines

checking bbb on xny
The Password Aging policy is==> DISABLED
The Account Life time ==> NEVER EXPIRES
Number of unsuccesful logins Allowed ==> INFINITE

checking qwe on qwqw
The Password Aging policy is==> DISABLED
The Account Life time ==> NEVER EXPIRES
Number of unsuccesful logins Allowed ==> INFINITE

checking eqw on inqw
The Password Aging policy is==> DISABLED
The Account Life time ==> NEVER EXPIRES
Number of unsuccesful logins Allowed ==> DEFAULT

checking eqw on inqw
The Password Aging policy is==> DISABLED
The Account Life time ==> NEVER EXPIRES
Number of unsuccesful logins Allowed ==> INFINITE

for the one that has DEFAULT i want to print out the 4 lines above that

checking eqw on inqw
The Password Aging policy is==> DISABLED
The Account Life time ==> NEVER EXPIRES
Number of unsuccesful logins Allowed ==> DEFAULT

how can i do this?

Thanks in advance

Re: grep help

Pete Randall — Tue, 28 Jul 2009 09:52:10 GMT

George,

From "Handy One-Liners for Sed" (attached):

# print paragraph if it contains AAA (blank lines separate paragraphs)
# HHsed v1.5 must insert a 'G;' after 'x;' in the next 3 scripts below
sed -e '/./{H;$!d;}' -e 'x;/AAA/!d;'

Pete

Re: grep help

James R. Ferguson — Tue, 28 Jul 2009 10:37:47 GMT

Hi :

You could do :

# perl -e '$/="";while (<>) {print if /DEFAULT/}' file

or:

# awk 'BEGIN{RS=""};/DEFAULT/ {print}' file

Regards!

...JRF...

Re: grep help

Dennis Handly — Tue, 28 Jul 2009 16:03:37 GMT

>JRF: # perl -e '$/="";while (<>) {print if /DEFAULT/}' file
># awk 'BEGIN{RS=""};/DEFAULT/ {print}' file

I'm not sure how these print out the previous 3 lines and the line with DEFAULT?
I'm assuming the files have zillions of lines and possibly a few of these 4 line blocks.

Of course here GNU grep would work with -B3.

Re: grep help

James R. Ferguson — Tue, 28 Jul 2009 16:15:55 GMT

Hi:

> Dennis: I'm not sure how these print out the previous 3 lines and the line with DEFAULT?

They do _IF_ as the post suggests that the input file is composed of paragraphs with a blank line as the stanza seperator.

Of course, that's my assumption and it would be obviously better if the OP had _attached_ the input.

...JRF...

Re: grep help

George Chechakunnil — Wed, 29 Jul 2009 03:47:05 GMT

Thanks for the Help but i am not getting the output as i want

The attached is the input file.. they in paragraphs each of 4 lines each.

Now i want the output to have all the paragraphs that have the word DEFAULT in it.

Desired output

checking kcrpw on inuap700
The Password Aging policy is ==> DEFAULT
The Account Life time ==> NEVER EXPIRES
Number of unsuccesful logins Allowed ==>DEFAULT(6)

checking kcrpw on inuap705
The Password Aging policy is ==> DEFAULT
The Account Life time ==> NEVER EXPIRES
Number of unsuccesful logins Allowed ==>DEFAULT(6)

checking glprd on inuap706
The Password Aging policy is==> DISABLED
The Account Life time ==> NEVER EXPIRES
Number of unsuccesful logins Allowed ==>DEFAULT(6)

etc etc

Thanks a lot in advance

Re: grep help

Dennis Handly — Wed, 29 Jul 2009 05:23:38 GMT

>The attached is the input file. they are in paragraphs each of 4 lines each.

Your input file doesn't have empty lines as separators, they have one space.

You will have to change what was given to:
sed -e '/...*/{H;$!d;}' -e 'x;/DEFAULT/!d' file

sed -e 's/^ *$//' file | awk '
BEGIN { RS=""}
/DEFAULT/ {print}'

Or explicit programming for what's there:
awk '
/^checking / {
getline l2; getline l3; getline l4
if (l2 ~ /DEFAULT/ || l3 ~ /DEFAULT/ || l4 ~ /DEFAULT/) {
print $0; print l2; print l3; print l4
}
}' file

>JRF: input file is composed of paragraphs with a blank line as the separator.

I found the standard where it describes RS="". It sure isn't on the man page.

Re: grep help

George Chechakunnil — Wed, 29 Jul 2009 08:28:56 GMT

Terrific

Thanks to all..

Re: grep help

Tingli — Fri, 31 Jul 2009 17:51:10 GMT

Although this post is too late, but I cannot help it.
sed -n -e ''/DEFAULT/'!{H;x;s/^.*\n$.*\n.*\n.*$$/\1/;x;}' -e ''/DEFAULT/'{H;x;p;}' File_Name

Re: grep help

Hein van den Heuvel — Fri, 31 Jul 2009 19:00:11 GMT

Also too late, also redundant, but maybe helpful to someone some day as it is more 'open ended'.

$ perl -ne 'if (/\S/) { push @p,$_ } else {print @p if grep /DEFAULT/,@p; @p=("\n")}' x.txt

This looks for (arbitrary) paragraph markers, here a line with only whitespace.

IF a line with non=whitespace is found, it is pushed onto a paragraph array p.
ELSE the array p that was build up is grepped for a target (DEFAULT) and the whole array printed on match. The array is then reset to a single new-line for the next go around.

btw... if there might not be a final paragraph marker, then you may want to turn it arouns some to catch a hit on the last paragraph:

perl -ne '{ push @p,$_ } if (/^\s*$/ or eof) {print @p if grep /DEFAULT/,@p; @p=()}' a.txt

fwiw,
Hein.

Re: grep help

James R. Ferguson — Fri, 31 Jul 2009 20:09:11 GMT

Hi (again):

I just noticed this thread bubble to the top again.

> Dennis: Your input file doesn't have empty lines as separators, they have one space.

Ahh...now I see the reason for the confusion and the failure of my original suggestion to yield the output that George wanted [read on].

> Dennis: I found the standard where it describes RS="". It sure isn't on the man page.

Ahh, yes, again!

Based on the original post (before the attachment of text), it appeared to me that blank lines consisted solely of newline characters and these separated the stanzas or paragraphs. My suggest was based upon that assumption.

In Perl, when that is the case, if you set the record separator ('$/') to a null string, one or more blank lines constitute a record terminator. Unfortunately, a "blanK line can consist only of a newline character without hidden spaces or tabs. Perhaps more unfortunately, Perl doesn't allow a pattern in place of a simple string for the record separator.

In the GNU 'awk' documentation, it is noted that an empty string for the value of 'RS' signals that records are separated by one or more blank lines. Dennis is correct; this _isn't_ described on the HP-UX manpages. It works, though.

Of my two original code snippets, I should have slightly modified the 'awk' variation to give back at least one blank line. The Perl snippet could be better written for the way the Forum blends whitespace too:

# perl -e '$/=qq();while (<>) {print if /DEFAULT/}' file

or:

# awk 'BEGIN{RS=""};/DEFAULT/ {print $0, "\n"}' file

Finally, to filter your raw data so that it is suitable to being piped to either of the above snippets, you could do (for example):

# perl -pe 's/^\s+$/\n/' file | perl -e 'BEGIN{$/=qq()};while (<>) {print if /DEFAULT/}'

Regards!

...JRF...