Closed daxiguaA666 closed 4 weeks ago
seems the serialversionUID of org.apache.iceberg.BaseFile
did not match, could you please check the serialversionUID
of org.apache.iceberg.BaseFile
in iceberg 1.4.2 and 1.0.0 jar file
The BaseFile of iceberg1.0.0 cannot be viewed, but I cleared the database, rebuilt the table, and used iceberg 1.4.2 to write it. The rewritten file serialization problem still exists.
OK, the problem is solved, the reason is that there is an additional iceberg dependency in the spark jar directory
What happened?
deserialization error when merging small files using the spark engine![企业微信截图_1716966460202](https://github.com/apache/amoro/assets/153582103/257b5212-5957-457a-8ee8-2f161100b90f)
Affects Versions
master
What engines are you seeing the problem on?
Spark
How to reproduce
The iceberg version is 1.4.2. The spark engine ran normally for a long time until data was written to a table using iceberg1.0.0. After that, all merge tasks failed. All tables were deleted, and a new iceberg1.4.2 table was created and used to write data. The merge task still failed.
Relevant log output
Anything else
The iceberg version is 1.4.2. The spark engine ran normally for a long time until data was written to a table using iceberg1.0.0. After that, all merge tasks failed. All tables were deleted, and a new iceberg1.4.2 table was created and used to write data. The merge task still failed.
Are you willing to submit a PR?
Code of Conduct