[SPARK-45655] current_date() not supported in Streaming Query Observed metrics

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.4.1, 3.5.0
• Fix Version/s: 4.0.0
• Component/s: Structured Streaming
• Labels: pull-request-available
• Assignee: Bhuwan Sahni
• Reporter: Bhuwan Sahni
• Created: 24/Oct/23 20:48
• Updated: 12/Nov/23 21:44
• Resolved: 12/Nov/23 07:50

Description

Streaming queries do not support current_date() inside CollectMetrics. The primary reason is that current_date() (resolves to CurrentBatchTimestamp) is marked as non-deterministic. However, current_date and current_timestamp are both deterministic today, and current_batch_timestamp should be the same.

 

As an example, the query below fails due to observe call on the DataFrame.

 

val inputData = MemoryStream[Timestamp]

inputData.toDF()
      .filter("value < current_date()")
      .observe("metrics", count(expr("value >= current_date()")).alias("dropped"))
      .writeStream
      .queryName("ts_metrics_test")
      .format("memory")
      .outputMode("append")
      .start()

 

Issue Links

• links to: GitHub Pull Request #43517