Hi Dan, I'm not very familiar with how Solr implements function-queries, but I agree this seems like a bug.
Taking a glance at the code it looks like most of the "math" function-queries cast to floats, so I wonder whether this behavior (intentional or not) wouldn't occur more widely too. e.g. "sum", "sub", "pow" all cast to float as well - would they be affected similarly by the cast-related roundin that you're running into for "mod"? Anyway, I'd say file a JIRA ticket for the issue. That should get more eyes on the problem and maybe someone with more context can correct my thoughts above. If you're willing to write up a patch/PR too I'd be happy to review! Best, Jason On Thu, Aug 18, 2022 at 12:35 PM Dan Rosher <[email protected]> wrote: > > Hi, > > I was surprised when using the mod function query, the modulus was wrong on > my largeish int. > > This was because the int was cast to float, meaning the modulus was wrong > for any int above 2^24 > > Should the mod function change to int modulus instead of float and have > another fmod function (like c/c++)? > > Dan --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
