[SPARK-25497] limit operation within whole stage codegen should not consume all the inputs

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.4.0
• Fix Version/s: 3.0.0
• Component/s: SQL
• Assignee: Wenchen Fan
• Reporter: Wenchen Fan
• Created: 21/Sep/18 02:36
• Updated: 05/Jan/19 15:28
• Resolved: 09/Oct/18 07:51

Description

This issue was discovered during https://github.com/apache/spark/pull/21738 .

It turns out that limit is not whole-stage-codegened correctly and always consume all the inputs

Issue Links

• is duplicated by: SPARK-25597 SQL query with limit iterates the whole iterator when WholeStage code generation is enabled
• is duplicated by: SPARK-26280 Spark will read entire CSV file even when limit is used
• links to: [Github] Pull Request #22524 (viirya)
• links to: [Github] Pull Request #22630 (cloud-fan)
• links to: [Github] Pull Request #22630 (cloud-fan)