From: David Vallner <david@...>
Date: 2006-12-03T03:15:54+09:00
Subject: Re: search and replace

--------------enigAF59EE22A0069F24937D8CA9
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

ishamid wrote:
> [total novice here]
>=20
> Hi,
>=20
> I have a series of expressions like this (shortened from verbose xml)
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> [<text:sequence text:ref-name=3D"refAutoNr0">1</text:sequence>
> [<text:sequence text:ref-name=3D"refAutoNr1">2</text:sequence>
> [<text:sequence text:ref-name=3D"refAutoNr2">3</text:sequence>
> [<text:sequence text:ref-name=3D"refAutoNr3">4</text:sequence>
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
>=20
> I want to globally replace each such line with just
>=20
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> \head
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
>=20
> followed by a line space so I get
>=20
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> \head
>=20
> \head
>=20
> \head
>=20
> \head
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
>=20
> etc.
>=20
> I am modifying a script with lines like
>=20
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> data.gsub!(/.*?<(office:text).*?>(.*?)<\/\1>.*/mois) do
>                 '\starttext' + "\n" + $2 + "\n" + '\stoptext'
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
>=20
> and don't yet know enough to completely understand. Probably a few more=

> hours/days of study will get me there but I need this urgently so...
>=20
> THNX in advance
>=20

Urght. *ducks*

> Best
> Idris
>=20
>=20

Regexps and XML always tend to blow up for me. The pattern you're
searching for seems to be a complete element, why not use <insert XML
parser of choice> and XPath?

With REXML, it should be something like:

document.elements.each('//text:sequence') {|sequence|
sequence.replace_with(REXML::Text.new("\\head\n", true))}

Substitute the XPath expression with one of desired precision. I'm a
little unsure around how REXML treats namespaces in XPath and such, but
if you know what prefix will be used in the document, that should work ou=
t.

The script might also require a little more massaging if you're
outputting to plaintext, but treating XML like, well, XML might get the
heavy lifting of searching for patterns in it done faster if you use a
pattern language operating on the DOM structure directly.

David Vallner


--------------enigAF59EE22A0069F24937D8CA9
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: OpenPGP digital signature
Content-Disposition: attachment; filename="signature.asc"

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (MingW32)

iD8DBQFFccLUy6MhrS8astoRAslhAJ45yqsdcH9HGm/SvN2MtjAV75M1VgCeN0Tx
UXiQ39eCXVG32nDKEQZjODI=
=vyHu
-----END PGP SIGNATURE-----

--------------enigAF59EE22A0069F24937D8CA9--