Why doesn't highlighting work on this document?

[Paul Rosen]

I have a puzzling case that I don't know how to begin to debug. I
present results with snippets highlighted, but it is not consistent, and
it would be nice to know why some documents are returned without any
highlighted text.

If you go to:

http://www.nines.org/search/saved?user=paul&name=tree

And look at the last entry on the page, (it should be titled "Nashe's
Red Herring: Epistemologies of the Commodity in Lenten Stuffe (1599)")

you'll see that there is no text returned for that object. I looked in
the solr index, and there is a fairly long text field, and in the middle
of that field is:

"which the tree, for example, maintains its form as a tree (wood
maintains itself in the specific form of the tree because this form is a
form of"

That is probably what was matched, but why wasn't that text returned?

(I'm using solr 1.4 nightly build from Sept 25)

[Jake Brownell]

By default the highlighter only considers the first 50k of text. See http://wiki.apache.org/solr/HighlightingParameters#hl.maxAnalyzedChars

Obviously the larger the size, the longer highlighting will take. When I get no highlights, I find that the text matched was later in the doc.

Jake

-----Original Message-----
From: Paul Rosen [mailto:paul@...]
Sent: Friday, November 06, 2009 4:07 PM
To: solr-user@...
Subject: Why doesn't highlighting work on this document?

I have a puzzling case that I don't know how to begin to debug. I
present results with snippets highlighted, but it is not consistent, and
it would be nice to know why some documents are returned without any
highlighted text.

If you go to:

http://www.nines.org/search/saved?user=paul&name=tree

And look at the last entry on the page, (it should be titled "Nashe's
Red Herring: Epistemologies of the Commodity in Lenten Stuffe (1599)")

you'll see that there is no text returned for that object. I looked in
the solr index, and there is a fairly long text field, and in the middle
of that field is:

"which the tree, for example, maintains its form as a tree (wood
maintains itself in the specific form of the tree because this form is a
form of"

That is probably what was matched, but why wasn't that text returned?

(I'm using solr 1.4 nightly build from Sept 25)

[Paul Rosen]

That was exactly the problem, thanks!

Jake Brownell wrote:

> By default the highlighter only considers the first 50k of text. See http://wiki.apache.org/solr/HighlightingParameters#hl.maxAnalyzedChars
>
> Obviously the larger the size, the longer highlighting will take. When I get no highlights, I find that the text matched was later in the doc.
>
> Jake
>
> -----Original Message-----
> From: Paul Rosen [mailto:paul@...]
> Sent: Friday, November 06, 2009 4:07 PM
> To: solr-user@...
> Subject: Why doesn't highlighting work on this document?
>
> I have a puzzling case that I don't know how to begin to debug. I
> present results with snippets highlighted, but it is not consistent, and
> it would be nice to know why some documents are returned without any
> highlighted text.
>
> If you go to:
>
> http://www.nines.org/search/saved?user=paul&name=tree
>
> And look at the last entry on the page, (it should be titled "Nashe's
> Red Herring: Epistemologies of the Commodity in Lenten Stuffe (1599)")
>
> you'll see that there is no text returned for that object. I looked in
> the solr index, and there is a fairly long text field, and in the middle
> of that field is:
>
> "which the tree, for example, maintains its form as a tree (wood
> maintains itself in the specific form of the tree because this form is a
> form of"
>
> That is probably what was matched, but why wasn't that text returned?
>
> (I'm using solr 1.4 nightly build from Sept 25)