An approach I've used successfully is to add a separate field called
'empty' and include the names of other fields that are empty, e.g,
empty: subject
You can search for;
+empty:subject
for messages with empty (null) subjects.
This approach means you don't need to think of a magical value that (you
hope) doesn't match other legitimate values. Additionally, you only add
one new field to your index for any number of null-able fields.
B.
Otis Gospodnetic wrote:
Hi George,
A field value needs to be indexed if you want to sort by it - see
http://www.lucenebook.com/search?query=index+for+sorting .
As for null fields, you could use an empty string, or better yet, some
"custom field value" that means "NULL" for your application.
Otis
--- George Abraham <[EMAIL PROTECTED]> wrote:
Hi,
I am using Lucene to index database content. I have two questions:
1. Along with some of the fields that I want indexed, there are some
fields that I want stored (not indexed) in the index so as to return
with hits. However I also want to be able to sort a search result
using these non-indexed fields. Which method should I use in the
Field
class then?
2. As is usual with some database content, sometimes fields that have
to be indexed turn out to be null. Using the Keyword method, the
indexing procedure gives me an error that 'value cannot be null'. How
do I then go about this?
Thanks,
George
____________________________________________________________________
Simpy -- simpy.com -- tags, social bookmarks, personal search engine
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
