Yes, I am.
The UID example Michael gave provides a way for us not to branch from
lucene code base.
I am trying to improve on it by storing the uid using position (since
position info is not used for ids) which would buy use in load time
quite a bit.
-John
On Nov 19, 2007 4:28 PM, Yonik Seeley <
On Nov 19, 2007 6:39 PM, John Wang <[EMAIL PROTECTED]> wrote:
> oh, is there a way of opening that?
Well, you can keep track of position increments yourself and then
choose the correct position increment so that the position you want is
indexed. AFAIK, positions increments must be positive , so y
oh, is there a way of opening that?
In the UID example Mike gave, it seems that uid can be stored in the
position part of the data.
It would be very efficient in both load time and index size to be able
to do that.
Thanks
-john
On Nov 19, 2007 1:22 PM, Yonik Seeley <[EMAIL PROTECTED]> wrote:
>
On Nov 19, 2007 4:14 PM, John Wang <[EMAIL PROTECTED]> wrote:
>What is the right way of setting customized position value on a
> token at indexing time.
You set the positionIncrement, and the lucene indexing code determines
the absolute position. You can't set an absolute position yourself.
Hi:
What is the right way of setting customized position value on a
token at indexing time.
Thanks
-John
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
