sohami commented on a change in pull request #1451: DRILL-6710: Disallow negative scale for VarDecimal URL: https://github.com/apache/drill/pull/1451#discussion_r213928418
########## File path: exec/java-exec/src/main/codegen/templates/Decimal/DecimalFunctions.java ########## @@ -170,6 +170,8 @@ public void eval() { left.precision, left.scale, right.precision, right.scale); + // Do the verification of computed scale and precision here based on actual data + typeInference.verifyScaleAndPrecision(); Review comment: I don't see moving the verification check here rather than doing in planning time helps. Since for each addition call we still increase the precision by 1, hence for 39 such additions precision will go beyond max allowed 38 and scale will become -1, hence throwing the error. I think instead what we can do is while passing `MathContext` to BigDecimal pass `adjusted precision` and `Math.max(0, adjusted_scale).` This will ensure that negative scale is never passed in context and if the operation is not possible in BigDecimal under given MathContext, it should throw Exception which Drill can in turn throw. I tried with that [temp change](https://github.com/sohami/drill/commits/DRILL-6710-2) and all the queries shared in the JIRA works with that. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services