Parsing logs, need to keep lines matching two conditions [SOLVED]

graysky · 2013-07-13 14:47:28

I wrote a script that generates a 12 columns in a text file based on parsing my apache logs. I need to remove any line that match the following two conditions:

1) <TAB>200<TAB>
2) <TAB>206<TAB>

Where <TAB> is a literal tab as my other stuff reads tab deliminated files. The following grep works to match simply '200' or '206' but I can't get the grep syntax correct to match with the trailing and leading tabs.

grep -w '200\|206' parsed.log

Any help is appreciated.

EDIT: Finally!

grep -E $'\t200\t\|\t206\t' parsed.log

EDIT2: Rrrr... the above works under zsh but not under bash... what's my problem?

Last edited by graysky (2013-07-13 16:16:08)

thisoldman · 2013-07-13 15:18:38

Does this work:

grep -E '(\t200\t)|(\t206\t)' parsed.log

Edit: Wait, that matches a literal 't', doesn't it?

I think you need this:

grep -P '[\t]200[\t]|[\t]206[\t]' parsed.log

Last edited by thisoldman (2013-07-13 15:33:54)

tavianator · 2013-07-13 15:19:50

graysky · 2013-07-13 15:26:53

Thanks for the assistance. Will mark as solved.

...actually, my original thought was flawed. I actually need to preform this grep query within column #9 in the text file. Simply doing it on a line-by-line basis does not catch the errors. I need to some how mate this up with awk.

EDIT: Here is a sample of the file: http://repo-ck.com/PKG_source/test.txt

Last edited by graysky (2013-07-13 15:58:23)

karol · 2013-07-13 15:33:03

graysky, do both conditions have to be met at the same time? Is it 1) AND 2) or 1) OR 2)?

graysky · 2013-07-13 15:33:58

karol wrote:

graysky, do both conditions have to be met at the same time? Is it 1) AND 2) or 1) OR 2)?

Is it an or: either 200 or 206 (it will never be both). I edited post #4 with a sample file.

Last edited by graysky (2013-07-13 15:48:25)

karol · 2013-07-13 16:03:17

graysky wrote:

Thanks for the assistance. Will mark as solved.
...actually, my original thought was flawed. I actually need to preform this grep query within column #9 in the text file. Simply doing it on a line-by-line basis does not catch the errors. I need to some how mate this up with awk.
EDIT: Here is a sample of the file: http://repo-ck.com/PKG_source/test.txt

Try

awk '$9 ~ /20[0,6]/' test.txt

The sample has only 200 in the 9th column, but I edited it and it worked with 206 too.

karol · 2013-07-13 16:08:21

Oooops.

You're using tab-deliminated files, so try

awk -F '\t' '$9 ~ /20[0,6]/' test.txt

instead.

Sorry for bumping.

graysky · 2013-07-13 16:15:39

@karol - Thank you. That solved it.

falconindy · 2013-07-15 13:14:56

Except that '/20[0,6]/' will also match "20," ... character ranges are literal.

Arch Linux

#1 2013-07-13 14:47:28

Parsing logs, need to keep lines matching two conditions [SOLVED]

#2 2013-07-13 15:18:38

Re: Parsing logs, need to keep lines matching two conditions [SOLVED]

#3 2013-07-13 15:19:50

Re: Parsing logs, need to keep lines matching two conditions [SOLVED]

#4 2013-07-13 15:26:53

Re: Parsing logs, need to keep lines matching two conditions [SOLVED]

#5 2013-07-13 15:33:03

Re: Parsing logs, need to keep lines matching two conditions [SOLVED]

#6 2013-07-13 15:33:58

Re: Parsing logs, need to keep lines matching two conditions [SOLVED]

#7 2013-07-13 16:03:17

Re: Parsing logs, need to keep lines matching two conditions [SOLVED]

#8 2013-07-13 16:08:21

Re: Parsing logs, need to keep lines matching two conditions [SOLVED]

#9 2013-07-13 16:15:39

Re: Parsing logs, need to keep lines matching two conditions [SOLVED]

#10 2013-07-15 13:14:56

Re: Parsing logs, need to keep lines matching two conditions [SOLVED]

Board footer