- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
02-08-2022 08:45 PM
Hi @Jose Gonzalez ,
Thanks for the reply.
Write stream is as follows
dataset.writeStream
.format("delta")
.outputMode("append")
.option("checkpointLocation", "dbfs:/checkpoints_v1/<table_name>")
.option("mergeSchema", "true")
.table("<table_name>")The Dataset is created after fetching records from delta tables in a stream and applying the flatMapGroupsWithState API on the records.
Few Observations:
1) The probability of the error occurring is more at the higher loads. If the input rate for the flatMapGroupsWithState API is 12000 records/second, then then the error occurs regularly and generally within 2 hours of the start of the job. For lower loads of 4000 records / seconds, the error occurs infrequently .
2) The size of the snapshot file for which the error occurs is 0 bytes.
Let me know if you require any other information.
Thanks & Regards,
Rohan