Github user mgaido91 commented on the issue: https://github.com/apache/spark/pull/22494 > If your argument is, picking a precise precision for literal is an individual featue and not related to #20023, I'm OK to create a new config for it. Yes this is - I think - a better option. Indeed, what I meant was this: let's imagine I am a Spark 2.3.0 user and I have `DECIMAL_OPERATIONS_ALLOW_PREC_LOSS` turned to `false`. Before this patch, I can successfully run `select 1234567891 / (1.1 * 2 * 2 * 2 * 2)`. After this patch, this query would return `null` instead, as an overflow would happen. So this patch is "correcting" a regression from 2.2 but it is introducing another one from 2.3.0-2.3.1. Using another config is therefore a better workaround IMO.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org