lucene-solr-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From okayndc <bodymo...@gmail.com>
Subject Re: escaping HTML tags within XML file
Date Sun, 25 Sep 2011 21:06:21 GMT
Here is a representation of the XML file...

<root>
<commenter>
<comment><p>Text here</p><img src="image.gif" /><p>More text
here....</p></comment>
</commenter>
</root>

I want to keep the HTML tags because it keeps the formatting (paragraph
tags, etc) intact for the output.  Seems like you're saying that the HTML
can be kept intact with the use of a HTML field type without having to
escape the HTML tags?

On Sun, Sep 25, 2011 at 2:52 PM, <pulkitsinghal@gmail.com> wrote:

> Assuming that the XML has the HTML as values inside fully formed tags like
> so:
> <node><HTML></HTML></node> then I think that using the "HTML"
field type in
> schema.xml for indexing/storing will allow you to do meaningful searches on
> the content of the HTML without getting confused by the HTML syntax itself.
>
> If you have absolutely no need for the entire stored HTML when presenting
> results to the user then stripping out the syntax at index time makes sense.
> This will adversely affect highlighting of  that document field as well so
> just know your requirements.
>
> If you don't want to present anything at all then don't store, just index
> and use the right field type (HTML) such that search results find the right
> document. Just because a field is helpful in finding the doc, doesn't mean
> folks always want to present it or store it.
>
> With Data Import Handler a HTML stripping transformer is present so that it
> is removed before the indexer gets it's hands on things. I can't be sure if
> that is how you get your data into Solr.
>
> - Pulkit
>
> Sent from my iPhone
>
> On Sep 25, 2011, at 8:00 AM, okayndc <bodymoves@gmail.com> wrote:
>
> > Hello,
> >
> > Was wondering if it is necessary to escape HTML tags within an XML file
> for
> > indexing?  If so, seems like a large XML files with tons of HTML tags
> could
> > get really messy (using CDATA).
> > Has this been your experience?  Do you escape the HTML tags? If so, what
> > technique do you use? Or do you leave the HTML tags in place without
> > escaping them?
> >
> > Thanks!
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message