I'm using the completion suggester with weights. It appears that the
weights for each suggestion is being rounded down. For example, if I have
suggestions "a" with weight 3.54 and "b" with weight 3.25, the weights for
both are rounded down to 3.0 when used for sorting and when they're
returned in the response.
Is this supposed to be the case and is there a way to use the weights with
double or float precision?
I think it makes sense to actually throw an error if a non integer is being
indexed as weight in order to prevent confusion. Simplest solution for that
might be to multiply by 100 (or whatever factor needed) in your case on the
client side.
I'm using the completion suggester with weights. It appears that the
weights for each suggestion is being rounded down. For example, if I have
suggestions "a" with weight 3.54 and "b" with weight 3.25, the weights for
both are rounded down to 3.0 when used for sorting and when they're
returned in the response.
Is this supposed to be the case and is there a way to use the weights with
double or float precision?
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.