Github user paul-rogers commented on a diff in the pull request:
https://github.com/apache/drill/pull/1026#discussion_r150689236
--- Diff: exec/java-exec/src/main/resources/drill-module.conf ---
@@ -502,6 +502,8 @@ drill.exec.options: {
store.format: "parquet",
store.hive.optimize_scan_with_native_readers: false,
store.json.all_text_mode: false,
+ store.json.writer.non_numeric_numbers: false,
+ store.json.reader.non_numeric_numbers: false,
--- End diff --
OK. So, if the rest of Drill either does not support (or we don't know if
it supports) NaN and Inf, should we introduce the change here that potentially
leads to failures elsewhere?
Do we know if JDBC and ODBC support these values? (I suppose they do as
they are features of Java's float/double primitives...)
---
