[SOLVED]Problem character using sed.

[buteman]

I have a number of html files from which I am trying to create csv files. I have tried a lot of options but the content of the files do alter each month. In my search I came across the use of sed which I have never used before. Well I have made some progress having seen a number of examples but have hit a wall with one part I am trying to alter.
I cannot work out the syntax for removing a piece of text and I think it's due to one of the characters.

<td style='border:1px solid black

I believe it's the ' character
I tried splitting it up as I thought I could escape it with \ but it doesn't work.

Code: Select all

cat ab.ht |sed 's/\'border//Ig'> ab.l

> '
sed: -e expression #1, char 14: unknown option to `s'
 

For some reason the complete section is

<td style='border:1px solid black>

but the final > appears in the next cell when viewed from Libreoffice calc

[xenopeek]

That doesn't work because single quoting in the shell reads all characters literally; no escape codes.

The alternative is to use double quoting so you can put a literal single quote in it:
sed "s/'border//Ig"

Or to use ANSI C quoting so you can use the \' escape sequence you used:
sed $'s/\'border//Ig'

Or you mix single and double quoting:
sed 's/'"'"'border//Ig'

See this section of the manual for explanations: https://www.gnu.org/software/bash/manua ... ml#Quoting

[buteman]

Thanks xenopeek did double quoting and that fixed it