| |
| |
| |
|
Comments:
<0> pkrumins: // means "use the last used pattern"
<0> Hence: sed '/foo/!d;s//bar/g' shows only lines matching /foo/ while replacing all occurences of foo with bar.
<1> how would i do a non greedy grab of everything between xyz and the very next semicolon?
<1> i'm looking for something like sed -e 's/xyz.*;/FOOBAR/', but the asterisk is taking everything up to the last ';' in the line
<2> hi all, I want to run a script with switches (like sh script param1 param2), within the script I need to use sed to replace some strings by 2nd switch, I've tryed sed -e 's/.*/&'$2'/', but don't woks, how can I do that?
<3> anything inside '' is taken literally.
<3> "s/.*/&$2/" , or 's/.*/&'"$2"'/' should work.
<3> although i read somewhere using "s///" is bad. dunno why
<2> goldfish: thanks
<2> Rado: thanks
<4> hi. I see this in a script sed -n -e "s#^$RPM_BUILD_ROOT##p"
<4> what does the '#' symbol do?
<4> shouldn't it be '/'?
<5> it is instead of /
<5> you can use "whatever" character instead of /
<4> so does the following cmd remove every line that has $RPM_BUILD_ROOT
<3> nope
<5> it is used when you want to use / in substitution, and you don't want to make it clumsy with all those
<5> \/
<3> well
<3> use "d" if you want to actually DELETE
<3> sed "/^$RPM_BUILD_ROOT/d"
<4> so why two '##'? doesn't that mean to replace refernces of $RPM_BUILD_ROOT with nothing
<5> it doesn't matter whether there is / or #
<3> adp: it does. but it wont delete the whole line
<4> goldfish: ok. got it.
<2> please give a rope, I need to p*** the 2nd parameter of a script ( $2 ) to sed to it replace a string by the value of the parameter, I can not find how to do that...
<6> Dantix: what do you have as the command right now?
<2> rawtaz_: sh script param1 param2
<6> inside the script then? the sed line
<2> rawtaz_: s/.*/&{$1}/
<2> rawtaz_: sorry, that is wht my sed command file contains
<3> hmmmmm.
<2> rawtaz_: I'm calling sed this way: sed -f sed.cmd
<3> (***ume bash) #!/bin/bash sed "s/string/$2/" file
<2> goldfish: that replaces string by $2 not by the value of the shell variable, I've just tested that
<3> it does not.
<3> you probably have it inside single quotes.
<3> not double quotes.
<3> In which case it would replace it with a literal $2, as anything inside '' is taken literally.
<2> goldfish: let me test it
<2> goldfish: this is my script: cat $2|cut -f1|sed 's/.*/&$1/' and I'm running: sh script @foo.com test.txt and it replaces the string matched by "cut" by $1 not by @foo.com
<3> yes.
<3> You just ignored what I said.
<3> It's inside single quotes, anything inside single quotes is taken literally, use double quotes.
<6> indeed
<3> also, no need for cat, sed can work on files, and you can most likely get rid of the cut. and just use sed.
<2> goldfish: mmm, sorry, perhaps my bad english...I'll try again
<3> sed "s/.*/&$1"
<2> goldfish: works!, thanks you very much by your help and patience :)
<3> No problem.
<3> cat $2| cut -f1| sed .... can be replaced with "cut -f1 $2| sed ...."
<3> Although, as I said, you can do it all in sed, but i think we'll leave that for another day's learning.
<2> goldfish: thanks again
<3> np.
<7> Hmm, what does this regex match? /^\([ -~]*\n\).*\n\1/
<7> Particularly [ -~]* is not clear to me
<3> a space, a -, or a ~, zero or more times.
<7> nope
<3> nope?
<7> i exchange [ ~-]*
<7> and it no longer produces the same resut as [ -~]*
<7> any ideas?
<7> goldfish, btw, talking about the yesterday's sed expression here is the answer: http://groups.google.com/group/comp.unix.shell/browse_frm/thread/b3885f5692585f64/1b43f015da0dfec3#1b43f015da0dfec3
<7> I am now deciphering this sed expression: sed -n 'G; s/\n/&&/; /^\([ -~]*\n\).*\n\1/d; s/\n//; h; P'
<7> which does eliminate nonconsecutive duplicate lines
<3> jesus.
<3> Some complicated stuff right there :)
<7> not too complicated
<7> just the [ -~]*
<3> well, i meant the explanation of yesterday's stuff.
<7> yes, that was pretty tricky
<3> well, i'm an idiot.
<3> heh
<7> takes time to become sed master
<3> [a-z] matches a to z
<7> sure
<3> so [ -~] matches space to ~
<7> but [-az] matches -, a or z
<3> indeed.
<7> you think it matches space to ~?
<3> yes
<3> although i'm not sure what way that works.
<7> where is my ascii chart table
<3> ascii values maybe?
<3> man ascii
<3> has it :)
<7> wow :)
<7> didnt know
<7> ha
<7> pretty cool
<7> it matches all printable chars
<3> handy alright.
<7> then
<3> Oh rigt.
<7> space is 0x20
<7> and ~ is 0x7E (176 dec)
<7> pretty cool
<7> wow nice trick
<7> match all printables [ -~]
<7> and 177 dec is DEL which is no longer printable just after ~
<3> i see
<7> makes sense and i fully understand this expression now.
<7> thanks goldfish
<7> :)
<3> np.
<3> man 7 regex
<3> is a decent resource, also.
<7> wow pretty good
<7> you know some more valuable man pages?
<3> well. it depends on what you value.
<3> man man , is valuable :)
<3> the -k option of man, searches....
<3> That's helpful at times.
<7> great
<7> man woman :)
<7> unfortunately does not exist
Return to
#sed
or
Go to some related
logs:
mozilla sirefox
inetd xwin fedora
mounting /dev/root failed, dapper
suse_register fails
supremedestiny boot
what do the dell 600m keyboard lights mean
php6+$_get
gentoo warning unable to open an initial console
#egghelp
gettext not translating
|
|