shell script to mv lines from log file to other file

[depele]

Hello,

I am looking for a shell script to move all lines from a log file starting with date before1/7/2008 to another file and then gzip it.

any suggestions.

thanks

[Greyed]

Your sentence is unclear. Do the lines start with the date or the file? If the lines, what is the format of the line?

[depele]

1 file
multiple lines starting with date

mv all lines date before 1/7/2008

I am looking for a preview now.

grtz...

[depele]

will this line do to cp all files?

cat logfile.log | grep "2008-07-*" >> test.txt


thanks in advance ?

[ghostdog74]

show a sample of your file.

[depele]

2007-12-08 13:07:44,887 ************************OK
2007-12-08 13:07:44,887 ************************************************** ************************************************** *
2007-12-08 13:07:44,928 DEBUG ************************************************** ************
2007-12-08 13:07:44,928 DEBUG ************************************************** ****************
<Destination>***************</Destination>
<Origin Role="DEFAULT_ROLE" OrgUnit="***************">*************</Origin>
*****************

2007-12-08 13:07:44,929 ************************************************** *

...... and so on.

==> all dates has to be all dates until 2008-07-01


grep "2008-[0-7]-*" > text.txt ==> will not work.

[ghostdog74]

describe what you want clearly. It obviously did not work because you are grepping for 2008

[depele]

There are also 2008 entries in the file. So that is not the problem.

I have a file (structured as above).
I need to add all lines starting with 2008-01-01 till 2008-07-01 to old.log
and I need to do the same
with all lines 2008-07-02 to new.log



is it clear now?

thank you

[geirha]

Try with
egrep '^(2008-0[1-6]-|2008-07-01).*'

[Greyed]

I have a file (structured as above).
I need to add all lines starting with 2008-01-01 till 2008-07-01 to old.log
and I need to do the same
with all lines 2008-07-02 to new.log


A file?



from datetime import date

infile = 'filename'
oldfile = 'filename.old'
newfile = 'filename.new'
div = date(2008,7,2)
log = open(infile,'r')
new = open(newfile,'w')
old = open(oldfile,'w')

for line in log:
words = line.split()
stamp = words[0].split('-')
filedate = date(stamp[0], stamp[1], stamp[2])
if filedate < div:
old.write(line)
else:
new.write(line)
Off-the-hip and untested but the gist is there. Python, of course, not shell. Basically, open the files needed, set a date object to the dividing line, read the lines in, parse out the date, construct a date object from it, compare the timestamp of the line to the dividing line, older than it write it to the old file, newer write to the new file. Could be shorter but wouldn't be as clear.

However if you're really needing it in shell for who-knows-what reason I defer to ghostdog74. My answer to "shell script" is Python. ;)

[depele]

thanks a lot.

I'm working on it.

grtz

[ghostdog74]

However if you're really needing it in shell for who-knows-what reason



egrep '^(2008-0[1-6]-|2008-07-01).*'


there's the reason : short and simple vs long winded. :)

[skotos]

If you used logrotate even with the files you are generating, the system would simply manage the rotation for you.:guitar:

Thus you might simply set a cron job - a bash one might be good enough - and complete the work, day by day, by taking away the rotated files, automatically compressed, if you need them to be compressed, gzipped, tarred or whatever.
:popcorn:

[depele]

If you used logrotate even with the files you are generating, the system would simply manage the rotation for you.:guitar:

Thus you might simply set a cron job - a bash one might be good enough - and complete the work, day by day, by taking away the rotated files, automatically compressed, if you need them to be compressed, gzipped, tarred or whatever.
:popcorn:

==> software not in my power to change logs everything in 1 file so the file keep on growing.
I already requested a change. But.........

it keeps on going

[Greyed]

there's the reason : short and simple vs long winded. :)

I could have used regex as well. I decided against it since split was perfectly servicable. Also that egrep magic isn't splitting like he requested.

[ghostdog74]

Also that egrep magic isn't splitting like he requested.

yes it sure does. Mind showing why it doesn't?

[Greyed]

Ye requested lines to go into TWO files. The egrep line, on its own, can only extract either lines that do match or do not match. Therefore you need two passes across the file to get the same results that my script does in a single pass. Also a single regex compile.

[ghostdog74]

Ye requested lines to go into TWO files. The egrep line, on its own, can only extract either lines that do match or do not match. Therefore you need two passes across the file to get the same results that my script does in a single pass. Also a single regex compile.
i see what you mean.


awk '/^(2008-0[1-6]-|2008-07-01)/{print >"testfile"}
/2008-07-02/{print> "nextfile"}' file