| |
| |
| |
|
Comments:
<0> How can I do a text incase sensative match .. .I thought it was like sed 'i/something/ah/g'
<1> hi,
<1> anyone know how i could use sed and write the output in the soruce file?
<2> sed -i 'commands' file
<3> mv FILE FILE.new; sed ..... FILE.new >FILE; rm -f FILE.new
<3> (which does the very same, GNU's -i does)
<1> ThBonsai, the pblem there is a lot of file and there are not in the same directory
<3> labreche: that's a problem of your shell, not of sed. you can't run sed recursively or similar. with or without -i
<3> both methods are equal. HOW sed gets its file arguments (or in the mv-case, how the loop around that gets its file arguments) is a question of your script/code/commands
<1> ok guys, tnx
<1> sed -i is the simpliest awy to do what i want
<1> TheBonsai, as i'm not really good with bash script, it will cost me too much time to make this script
<1> but i 'll try
<3> find <whatever> -exec <whatever> \; ?
<1> ThBonsais: for the recursive exec of sed -i, you got it: find -iname *.php|wargs sed -i command
<3> you mean xargs, and you can write find -exec sed -i {} +
<3> and you want to quote your pattern
<3> i mean: find -iname "*.php" -exec sed -i {} + should do it aswell (if your find doesn't lack the standard)
<3> regarding the + i mean
<1> Hi gnubin
<1> Hi gnubie
<1> Hi gnubien
<4> lol.
<1> i'm not well waed up ;-)
<5> labreche: hi
<4> labreche: use tab completion !
<1> tnx goldfish
<5> the tab is your easy typing spell checker best friend ;)
<4> :)
<6> is there a way to remove characters like <>#"/ in a text file?
<3> yes
<6> k
<6> but how
<6> cant get the job done
<4> Why not
<4> use the 's' command
<3> i wouldn't use sed 's/<>#"\///g' for that, i'd use tr -d '<>#"/'
<6> i got bad flag in substitute command
<4> use tr.
<4> as TheBonsai showed.
<6> k
<6> tr -d '<>#"/' < file does not work, i want to remove a exact line <>#"/
<4> hmmm?
<4> You stated you wanted to remove those chars in a file.
<4> sed '/^<>#"\/$/d'
<6> k
<6> i got invalid command code
<4> paste it
<6> goldfish, <>#"/ is just a example but the real line that must be remove a xx time in that html file is different ill past it ..
<4> ah jesus
<4> !!!!
<6> it is extualy a hole block of html code that must be removed several times
<4> ok
<4> pastebin the file
<6> www.pastebin.be/1787/
<4> and what do you want removed?
<6> line 2
<6> and those two section buttons
<4> hm.
<6> well exualy that table it is a ugly one
<6> i it possible?
<6> i/is
<6> line for line is allso okay
<4> Well.
<4> For that particular piece of text, sed '/^<td><font.*$/d;/<a.*$/d' , could work.
<4> but it's not a very good solution....
<6> hmm, but it could allso remove lines that dont must be removed?
<4> exactly
<6> could tr do the job
<4> nope
<4> you could do it in sed
<4> if you had the full file , you could hack up a sed command to do it
<6> what characters must be given to isolate line 2 for example
<4> the ones i gave are fine
<4> for that sample text
<4> it depends on what other <td> lines are in the file
<6> i could allso give in your example the complete exact line 2 including spaces ?
<4> you could
<6> k
<4> you wil have to escape /
<4> so replace each / with \/
<6> in that line, ok
<6> it does not work goldfish
<4> sed '/<td><font color="black" size=1>I l<font color="#FF0000">@<\/font>ve RuBoard<\/td>/d'
<4> works for me
<6> you are right! it works thanks ;)
<4> np.
<6> am wondering if it is possible without using \ for / and isolate the string and command sed to see the line litterly as exact line of characters?
<4> well
<4> usually you can use a different delimiter for sed
<4> i.e insstead of s///
<4> you can use s@@@
<4> or whatever
<4> it appears not to be the case with my sed , for the //d command
<4> so you will have to replace any / with \/ (i.e "escape it")
<6> i see
<6> so you can use s@@@ but not with combination with the delete command?
<4> well
<4> it appears you have to use //d
<4> [%] sed '@a@d' a
<4> sed: -e expression #1, char 1: unknown command: `@'
<3> \@
<4> hmm.
<4> TheBonsai: thanks.
<4> klauwhamer: sed '\@blahblahblahblha@d'
<4> or whatever delimiter you want to use.
<6> hmm
<6> cool
<6> but it cannot handle <>
<3> define "handle"
<3> \> and \< are word boundaries in some RE dialects. maybe that's where you collide with it
<6> so i dont get rit of that \
<3> ?
<3> /foo/ in general is no RE, it's a line address
<6> thebonsai, i mean if i use a exact string including some special characters like <> / " # , can i use that string without having to use these \ in that string?
<3> when \< is special, < is literal
<3> when * is special, \* is literal
<3> easy princip
<3> just don't let your shell eat some \\\\\
<3> use single quoting wherever you can
<6> so with single quonting i can isolate a string literal?
<3> completely literal, yes (POSIX shells, all bourne shells)
<3> the only recognized character is (the closing) '
<3> that also explains why you need a thing like 'foo'\''s bar' to get: foo's bar
<3> because the \ doesn't work inside '', no way to do 'foo\'s bar'
<6> interresting
<6> cool
<6> thanks
Return to
#sed
or
Go to some related
logs:
povaddict debian -site:boincstats.com
#css
#linux
xmms goodnight halt
ion3 on centos
install f4l suse 10.1
#lisp
#web
#ubuntu
/INCREMENTAL:NO + mozilla
|
|