ad1happy2go commented on issue #9032:
URL: https://github.com/apache/hudi/issues/9032#issuecomment-1607304355
@zaza I tried to reproduce the issue with hudi 0.13.1, but I am seeing the
appropriate behaviour only for bulk insert too. All CDC rows are coming as
inserts which is expected for bulk insert.
Can you let me know on exactly what scenario you are getting
NullPointerException? Is it intermittent?
Code I tried -
```
val path="file:///tmp/output/issue_9032_4"
val dataGen = new DataGenerator
val inserts = convertToStringList(dataGen.generateInserts(10))
val df = spark.read.json(spark.sparkContext.parallelize(inserts, 2))
val options = Map(
"hoodie.table.name" -> "line_items",
"hoodie.datasource.write.recordkey.field" -> "uuid",
"hoodie.datasource.write.precombine.field" -> "ts",
"hoodie.datasource.write.partitionpath.field" -> "partitionpath",
"hoodie.parquet.max.file.size" -> "125829120",
"hoodie.parquet.small.file.limit" -> "104857600",
"hoodie.index.type" -> "BLOOM",
"hoodie.bloom.index.use.metadata" -> "true",
"hoodie.cleaner.policy" -> "KEEP_LATEST_COMMITS",
"hoodie.cleaner.commits.retained" -> "168",
"hoodie.keep.min.commits" -> "173",
"hoodie.keep.max.commits" -> "174"
)
df.write.format("hudi").options(options).option(DataSourceWriteOptions.OPERATION.key,
"bulk_insert").option(HoodieTableConfig.NAME.key(), "line_items")
.option(HoodieTableConfig.CDC_ENABLED.key, "true")
.option(HoodieTableConfig.CDC_SUPPLEMENTAL_LOGGING_MODE.key,
HoodieCDCSupplementalLoggingMode.data_before_after.name())
.mode("append")
.save(path);
spark.readStream.format("hudi").option("hoodie.datasource.query.incremental.format",
"cdc").option("hoodie.datasource.query.type", "incremental")
.load(path)
.writeStream.foreachBatch { (batch: Dataset[Row], _:
Long) => batch.show(false); }.start.awaitTermination;
```
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
