Replace text with regular expression?

Question

I m looking for a way to find and replace large number of text files. For example;

I want to choose;

<li><a href="">Istanbul, TR POS </a></li>
<li><a href="">Ankara, TR POS </a></li>
<li><a href="">Izmir, TR POS </a></li>

WITH;

<li><a href="pos-istanbul-tr.php">Istanbul, TR POS </a></li>
<li><a href="pos-ankara-tr.php">Ankara, TR POS </a></li>
<li><a href="pos-izmir-tr.php">Izmir, TR POS </a></li>

Notice that initial of the label is lowercased and also added as part of the link. This should be done for a large number of text files so i m looking for the most efficient way via regex or any software that you think that might help.

Answer 1

I think you're going to need programming to do this, as you want to manipulate the matching text.

Sounds very doable with awk, if you're on a platform that has it. Or you could whip something up in Python, Perl, or whatever you prefer. There will very likely be other answers with actual code.

Answer 2

If you're on Linux you might find this thread helpful:

• Replace in multiple files - graphical tool for Linux (Regexxer http://regexxer.sourceforge.net/ was the accepted answer)

You could use a command line tool like sed, a scripting language like Python/Perl, or any number of other solutions to do this. If you can give more details as to what you're looking for and what OS it needs to run on that would help in providing a more specific answer.

Answer 3

Use some text editor capable of regexp and "search in files". E.g. EditPlus

then replace

<li><a href="">([A-Za-z]+), TR POS <\/a><\/li>

with

<li><a href="pos-\1-tr.php">\1, TR POS </a></li>

(Might need some more escaping, i.e. backslashes...)

Answer 4

$ perl -i.bak \
>  -pe's/href="">([^,]+)/\'href="pos-\'. (lc $1) . \'-tr.php">\'. $1/eg' \
>  *.html

Cross-platform variant (+ building on @Jonathan Leffler's answer)

Save it to fill-href.pl:

#!/usr/bin/perl -w -pi.bak
s/href="">([^,]+)/href="pos-\L$1\E-tr.php">$1/g

Run:

perl fill-href.pl test1.html test2.html

Answer 5

perl -pi.bak -e 's%<li><a href="">(\W+), TR POS </a></li>%<li><a href="pos-\L$1\E-tr.php">$1, TR POS </a></li>%g;' file1 file2 ...

Untested - probably over simplified, but should work on sample data. The '-p' causes Perl to print each line; the '-i.bak' creates a backup of the file with '.bak' extension and overwrites the original.

Answer 6

If you happen to have access to a Perl compatible regex (PCRE) engine, for example PHP's preg_replace(), or even Perl, if you must ;-), you can replace this regex:

<a href="">([^,]+),\s+(\w+)\s+(\w+)

with this:

<a href="\L$3-$1-$2\E.php">$1, $2 $3

The \L and \E modifiers do the lower-casing for you.

$i = '<li><a href="">Izmir, TR POS </a></li>';
$r = '/<a href="">([^,]+),\\s+(\\w+)\\s+(\\w+)/';
$s = '<a href="\\L$4-$2-$3\\E.php">$1, $2 $3';

$o = preg_replace($r, $s, $i);
echo $o;

VI / VIM have a similar mechanism of manipulating case in back-references.

Answer 7

Using textmate regex engine what you need is this:

find: <li><a href="">([A-Za-z]+), TR POS <\/a><\/li>

replace: <li><a href="pos-\L$1-tr.php">$1, TR POS </a></li>

Observe that the first replacement has a modifier that takes the first match to lower case. The second $1 doesnt have the modifier. I test it and it does exactly what you need.