umehrot2 commented on a change in pull request #1330: [HUDI-607] Fix to allow
creation/syncing of Hive tables partitioned by Date type columns
URL: https://github.com/apache/incubator-hudi/pull/1330#discussion_r380339707
##########
File path: hudi-spark/src/main/java/org/apache/hudi/DataSourceUtils.java
##########
@@ -77,6 +80,11 @@ public static Object getNestedFieldVal(GenericRecord
record, String fieldName, b
// return, if last part of name
if (i == parts.length - 1) {
+
+ if (isLogicalTypeDate(valueNode, part)) {
Review comment:
This method seems to me like the right central place to perform this check
and conversion. If we check the accessors of this method, it is used by Hudi to
retrieve the key values for its metadata fields. It might be more standard to
have Hudi treat Date as the actual Date string instead of a Long, across all
its keys for its internal usage. It would create a lot of confusion otherwise
and accessors of this functions will have to take care of adding this check.
Also, we would again have to re-write the same logic/loop again on client
side to check for logical type. Because this function basically returns only
the value.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
