[SPARK-23062] EXCEPT documentation should make it clear that it's EXCEPT DISTINCT

• Type: Improvement
• Status: Resolved
• Priority: Trivial
• Resolution: Fixed
• Affects Version/s: 2.0.0
• Fix Version/s: 2.3.0
• Component/s: Documentation, SQL
• Assignee: Henry Robinson
• Reporter: Henry Robinson
• Created: 13/Jan/18 00:04
• Updated: 17/Jan/18 08:03
• Resolved: 17/Jan/18 08:03

Description

Between 1.6 and 2.0, the default behaviour for EXCEPT changed from EXCEPT ALL to EXCEPT DISTINCT. This is reasonable: postgres defaults to the same, and if you try and explicitly use EXCEPT ALL in 2.0 you get a good error message. However, the change was confusing to some users, so it's worth being explicit in the documentation.

Issue Links

• links to: [Github] Pull Request #20254 (henryr)