When one search term started bombing, I discovered that one of my documents
somehow must be returning an NaN for one of the custom score scripts
fields. I was using the ternary operator already for emptiness:
I've tried using variations on the MVEL "is" and "instanceOf" decimal but I
don't know if or how those are supported as every variation throws an
error.
have you tried
doc['custom_booster'].value instanceof Number
You may want to think about index time boosting (although not that
flexible) if you are using this script all the time.
--Alex
On Tue, May 14, 2013 at 3:00 PM, Justin Treher jtreher@gmail.com wrote:
When one search term started bombing, I discovered that one of my
documents somehow must be returning an NaN for one of the custom score
scripts fields. I was using the ternary operator already for emptiness:
I've tried using variations on the MVEL "is" and "instanceOf" decimal but
I don't know if or how those are supported as every variation throws an
error.
What type is your 'custom_booster' mapped as? If it is a number, I'd be
interested to see what value is producing a NaN. Any chance you can
reproduce?
clint
On Tue, May 14, 2013 at 3:40 PM, Alexander Reelsen alr@spinscale.de wrote:
Hey
have you tried
doc['custom_booster'].value instanceof Number
You may want to think about index time boosting (although not that
flexible) if you are using this script all the time.
--Alex
On Tue, May 14, 2013 at 3:00 PM, Justin Treher jtreher@gmail.com wrote:
When one search term started bombing, I discovered that one of my
documents somehow must be returning an NaN for one of the custom score
scripts fields. I was using the ternary operator already for emptiness:
I've tried using variations on the MVEL "is" and "instanceOf" decimal but
I don't know if or how those are supported as every variation throws an
error.
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.