[Bug] paimon 0.8 spark read tag error about paimon

[Zouxxyy]

I haven't reproduced it. Can you provide more detailed steps to reproduce it, is the tag written in spark?

[eric666666]

I haven't reproduced it. Can you provide more detailed steps to reproduce it, is the tag written in spark?

I find the reson. The tag is created by flink at paimon 0.8 snapshot version which compliled early than 2024-04-15, But after 2024-04-15 Tag class add two new class field.
My tag file content:

{
  "version" : 3,
  "id" : 1,
  "schemaId" : 0,
  "baseManifestList" : "manifest-list-0dfb8ec3-0445-45f9-9c59-4d08e6443c43-0",
  "deltaManifestList" : "manifest-list-0dfb8ec3-0445-45f9-9c59-4d08e6443c43-1",
  "changelogManifestList" : "manifest-list-0dfb8ec3-0445-45f9-9c59-4d08e6443c43-2",
  "commitUser" : "59c08f33-b5de-4478-823c-5e681ee6e0b3",
  "commitIdentifier" : 1,
  "commitKind" : "APPEND",
  "timeMillis" : 1714379998639,
  "logOffsets" : { },
  "totalRecordCount" : 2531,
  "deltaRecordCount" : 2531,
  "changelogRecordCount" : 8227,
  "watermark" : 1714407456182
}

However new commit add new class fields:

Spark read tag system table‘s jar is compiled after 2024-04-15.
So new version code read tag's tagCreateTime using LocalDateTime.MIN.

This directly made next exception

Caused by: java.lang.ArithmeticException: long overflow
	at java.lang.Math.multiplyExact(Math.java:892)
	at org.apache.spark.sql.catalyst.util.DateTimeUtils$.millisToMicros(DateTimeUtils.scala:240)
	at org.apache.spark.sql.catalyst.util.DateTimeUtils$.fromJavaTimestamp(DateTimeUtils.scala:190)
	at org.apache.spark.sql.catalyst.util.DateTimeUtils.fromJavaTimestamp(DateTimeUtils.scala)

Because LocalDateTime.MIN timestamp millis convert to micros is bigger than Long.MAX_VALUE.

I think there could use a big Date to instead LocalDateTime.MIN which is just a symbol to express old Tag file do not having tagCreateTime field. Such as we could use Epoch time.
So i will submit a PR to fix it.

[JingsongLi]

Thanks @eric666666 for reporting and investigation. We should return null for these two fields when they are nulls.