bash script to apply sed command only to a specific text area

mauran · 07-13-2007, 01:28 PM

Dear all,

I need to apply an sed argument to specific text area in a html file.

first need to find the pattern <font=xxx> 

then apply sed argument within the font tag only.

can any one help me to write this bash script?

acid_kewpie · 07-13-2007, 01:36 PM

well that's the point of sed...

sed 's/\<font color\=red/\<font color\=blue/g' file.txt

should do it as an example.

mauran · 07-13-2007, 02:05 PM

Quote:

Originally Posted by acid_kewpie

well that's the point of sed...

sed 's/\<font color\=red/\<font color\=blue/g' file.txt

should do it as an example.

sorry.

My need is not just changing the font tag. I don't want to change it. I need to apply a sed argument to the text which coming within the specific font tag.

The argument should not be applied to the other text in that html file.

pixellany · 07-13-2007, 02:38 PM

Quote:

Originally Posted by mauran

sorry.

My need is not just changing the font tag. I don't want to change it. I need to apply a sed argument to the text which coming within the specific font tag.

The argument should not be applied to the other text in that html file.

He just gave an example.....

The sed replace (s) command will change the 1st instance or, with the "g" added, ALL instances. I seem to recall that there is a way to select a specific number of instances.

My favorite tutorial on sed is here: http://www.grymoire.com/Unix/Sed.html

It will help to post an example of text that you want to change.

acid_kewpie · 07-13-2007, 02:39 PM

that's still just a standard sed command, not sure what the issue is...

sed 's/\some\ data\ here/\some\ new\ data/g' file.txt

mauran · 07-13-2007, 02:47 PM

ok I'll explain it in detail.

here is the html file.

Quote:

<BODY LANG="en-US" DIR="LTR">
kl;lf;fsg;G
(baticoloa) rpj;jhz;b
Nfhusq;Nfzp gpuNjrj;jpy; jkpo; ,isQd; xUtu; milahsk; njupahj
MAjjhupfspdhy; Rl;Lf;nfhiy nra;ag;gl;Ls;shu;. Rl;Lf;nfhiy
nra;ag;gl;ltu;> rpj;jhz;b ehtyu; tPjpiar; Nru;e;j kJutPud; Rnu];
(suresh) (taJ 19) vd
njupate;Js;sJ. jdJ ez;gu;fSld; Js;Se;J xd;wpy; ,d;W tpahof;fpoik fhiy
6.30 kzpf;F nrd;W nfhz;bUe;jNghJ ,j;jhf;Fjy; elj;jg;gl;lJ. mtuJ
rlyj;ij kPl;fr;nrd;w NghJ rlyj;jpw;F mUfpy; Fz;L xd;W fhzg;gl;ljhf
rpwpyq;fh fhty;Jiwapdu; njuptpj;Js;du;. kuz tprhuizfspd; gpd;du; ,d;W
gpw;gfy; RNu]; rlyk; cwtpdu;fsplk; xg;gilf;fg;gl;lJ.
</BODY>

You can see there are two font varients.

1. &&&&&&&&&&&

2. **********

the script will search entire html, and will find the pattern **********

and then it will apply this sed command only to the text which situated by "**********" tags (the *****)

Quote:

sed -e 's/—/௮/g' -e 's/•/௬/g' -e 's/‚/ஸ்ரீ/g' < $filename > $filename-tmp-tsc2uni

important thing is, it should not make any changes to other text areas.

acid_kewpie · 07-13-2007, 04:38 PM

as above. that is what sed is for. this is not special in any way whatsoever...