| |
| |
| |
|
Comments:
<0> hi I'm just starting with sed, and I'm having difficulty removing a ' from an input line
<0> sed 's/\'//' is asking me for a matching '
<0> what am i doing wrong? sed s/'// doesn't work either
<0> um this is in a pipeline if that makes a difference?
<0> OK, I need a sed cmd to get rid of [, ] or ' (3 chars) in an input line- can anyone help please?
<1> sed s/"'"//g file
<0> thanks beebum, got same answer in bash
<0> glad you're awake tho ;)
<1> heh, i'm not awake ;-P
<2> igli: sed 's/[]['\'']//g'
<0> what you're sleep-typing? that's dangerous m8
<0> goedel: sed "s/[],'[]//g" which looks really funky
<0> (sorry added a ,)
<0> sed "s/[]'[]//g" looks mad but works
<0> goedel: i can't scan your version. are the extra ' chars to deal with ' as string delimiter?
<2> igli in '\'' the first quote ends the string before; \' is an escaped quote; and the third quote starts the next string
<2> igli: so you can embed a single quote in a single quotes delimited string in bash
<0> now ic - was wondering about interaction with bash, but it's all bash
<0> everything's a string
<0> if i wanted to delete strings of the form UNKNOWN_nn where n is a digit would that be sed s/UNKNOWN_[0-9][0-9]//g ?
<0> more to the point, how do i combine that with the previous cmd to execute in one invocation?
<0> sed "s/[]'[]//g" -e s/UNKNOWN_[0-9][0-9]//g ?
<0> try it and see..
<2> sed "s/[]'[]//g;s/UNKNOWN_[0-9][0-9]//g"
<0> lovely, ty goedel
<2> np igli
<0> :)
<3> hi
<4> yes
<3> I have lines like "memberUid: <variable_string>,ou=....." . I want to remove all starting from the "," (including it). How can sed manage that ?
<4> sed 's/,.*//'
<3> ho, it's that simple. Thanks
<3> but no
<4> elshaa: provided you dont have more than one , and you want to remove only from the second onwards
<3> no, the "memberUid: " in from of the line is important.
<3> almost all lines in the file contains ",.*"
<4> $ echo "memberUid: <variable_string>,ou=....." | sed 's/,.*//'
<4> memberUid: <variable_string>
<4> what are you looking for then
<5> and you only want to do it on lines containing memberUid: ?
<3> echo "stuff: <variable_string>,ou=....." | sed 's/,.*//'
<3> stuff: <variable_string>
<4> that is not what you need ?
<3> I only want lines with leading "memberUid" to be replaced
<4> ah
<4> goldfish: you are good at guessing :)
<5> oh yes!
<4> sed '/memberUid/s/,.*//'
<4> sed '/^memberUid:/s/,.*//' better
<3> thanks !
<4> specifiy your requirements completely. not everyone is good at guessing like goldfish :)
<3> ok :)
<6> hi, I'm just wondering about a sed command - here: http://dpaste.com/4895/ for some reason webalizer.conf is always empty after I run the script, and webalizer doesn't work (obviously) after this... what am I doing wrong?
<5> Use a temp file.
<5> sed 's///' file > tmpfile && mv tmpfile file
<5> also, for f in logs/access*; do ...; done
<5> don't use for f in $(ls)
<6> right, thanks
<7> genelisp: some seds suport -i
<1> that's very strange, i use webalizer off and on and have not seen that problem
<1> i'm curious to know what's causing it to wipe out the .conf file
<5> his redirection.
<1> ah
<1> just opened the link
<6> goldfish, ok, I changed it, and it works :-) but... it's appending 'logs/logfile' to the line - what modification could I make so it just adds 'logfile'?
<5> oh, you're using $f in the sed.
<6> yes
<5> ${f#*/}
<6> many thanks :-) I'll give it a go - I'm rubbish at bash :-S
<5> Or just hardcore it in.
<5> #bash -- where all your dreams come true
<8> oh my :)
<9> hey guys, i've got a log that i need to preprocess, i have a line like this:
<9> 10.1.5.170 - - [22/Jan/2007:14:07:01 -0700] "GET /blah HTTP/1.0" 200 25 "-" "Wget/1.10.1 (Red Hat modified)" "someip:53575 -> 10.1.5.170:80"
<9> and i want to get the 'someip' part at the front
<9> and replace that 10.1.5.170 at the front
<9> any idea?
<9> sed 's/\(.*\) \(.*\) \(.*\) \(.*\)/\2 \1 \4 \3/' sort of does the trick lol
<9> quite a hack
<8> i would have done a similar thing, maybe anchored to $
<8> prec?
<9> sed 's/\(.*\) "\(.*\):\(.*\) \(.*\)/\2 \1 \4 \3/' is good enough
<9> prec ?
<8> a person
<8> well, maybe not a person, more a semi-sed-god
<8> "semi" - bad work
<8> "half"
<0> demi-god is the expr i believe
Return to
#sed
or
Go to some related
logs:
unbuntu install amule
#qemu
debian dsp no such deivce
#python
Cannot modify header information makeThumbnail
postcat queue
#centos
#osdev
set_app_type
#gentoo
|
|