Adding a line to multiple files using Sed

Tuxqi · 11-26-2009, 05:09 PM

Ok, I have a load of php/html files all structured the same way (about 90).

Each page currently ends with:

Code:

<hr />
</div>
</body>
</html>

but I want to add a new line between <hr /> and </div>

so it now looks like this:

Code:

<hr />
<?php include("../footer.php"); ?>
</div>
</body>
</html>

I suppose an easier way is to delete the last 3 lines and add them to the footer.php file. Altough I am unsure how to do that using sed as well.

ghostdog74 · 11-26-2009, 06:13 PM

The basic logic behind this is, get to the line where <hr /> is, print that line, then print your required line, then do the rest. Its that simple

Code:

$ awk '/<hr \/>/{print;print "new";next}1' file

in bash

Code:

while read -r line
do
  case "$line" in
     *"<hr />"*) 
        echo "$line"
        echo "my new line"        
        ;;
     *) echo "$line"
  esac
done < "file"

Tuxqi · 11-27-2009, 02:01 AM

problem with that is, that there are a few <hr />, the thing that distinguishes the one I want to add a line underneath is that </div> comes underneath this particular <hr />.

druuna · 11-27-2009, 02:17 AM

Hi,

This seems to be working:

sed '/<\/div>/,/<\/body>/s%$<\/div>$%<?php include("../footer.php"); ?>\n\1%' infile

Testrun:

Code:

$ cat infile
<hr />
<td>
</td>
<hr />
<br>
<asdfasdf />
<hr />

<hr />
</div>
</body>
</html>


$ sed '/<\/div>/,/<\/body>/s%\(<\/div>\)%<?php include("../footer.php"); ?>\n\1%' infile
<hr />
<td>
</td>
<hr />
<br>
<asdfasdf />
<hr />

<hr />
<?php include("../footer.php"); ?>
</div>
</body>
</html>

If this works for you you need to add the -i switch to sed (sed -i '/.......) to make the change in place. A simple loop will make sure you process all the files.

Hope this helps.

Tuxqi · 11-27-2009, 01:58 PM

That's nearly perfect but its added the line before every instance of the </div> tag.

druuna · 11-27-2009, 02:27 PM

Yes, that's where it should be if I look at example in post #1:

Quote:

but I want to add a new line between <hr /> and </div>

so it now looks like this:

Code:
<hr />
<?php include("../footer.php"); ?>
</div>
</body>
</html>

My output is the same as your requirement........

If your example isn't correct or your requirements are not as stated it is kinda hard to come up with something that works. Please elaborate what it is you actually want.

Tuxqi · 11-27-2009, 02:36 PM

Sorry for not being 100% clear,

what I was saying was that the part of the code is unique as there is only one part of it where the <hr /> appears and then the </div> comes next.

thats why I said between those two as its unique to the rest of the code.

druuna · 11-27-2009, 02:57 PM

Probably me, but I really don't get what it is you are trying to say.

Could you provide a relevant example of your input file? And at least include what triggered the false hits or other problems.

Tuxqi · 11-27-2009, 03:01 PM

this is what happened.

<br />
<?php include("../footer.php"); ?>
</div>
<hr />
blah blah blah text text blah blah blah
<hr />
<?php include("../footer.php"); ?>
</div>
</body>
</html>

It also added a line between <br /> and </div>, which is an unwanted result.

druuna · 11-27-2009, 03:23 PM

Ok, I get it

lineNo=$((`sed -n '$=' infile`-2)) ; sed "${lineNo}s%$<\/div>$%<?php include("../footer.php"); ?>\n\1%" infile

A nice one liner

What it does:

1) Determine the amount of lines in the file (sed -n '$=' infile),
2) subtract 2 (To get the linenumber where the last </div> tag is), store in vatiable lineNo
3) replace on that specific line (sed "${lineNo}s%$<\/div>$%<?php include("../footer.php"); ?>\n\1%" infile)

1 assumption: All pages end with </html>. If any empty/other lines are after that my solution will fail.

Not an elegant solution, but I hope it does what you want!