[SPARK-31790] cast scenarios may generate different results between Hive and Spark

• Type: Documentation
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 2.4.5
• Fix Version/s: 3.0.0
• Component/s: Documentation
• Assignee: philipse
• Reporter: philipse
• Created: 22/May/20 03:00
• Updated: 12/Dec/22 18:10
• Resolved: 22/May/20 13:04

Description

`CAST(n,TIMESTAMPTYPE)` If n is Byte/Short/Int/Long data type, Hive treat n as milliseconds unit , while Spark SQL as seconds unit. so the cast result is different,please be care when you use it.

For example:

In spark
spark-sql> select cast(1586318188000 as timestamp);
52238-06-04 13:06:400.0
spark-sql> select cast(1586318188 as timestamp);
2020-04-08 11:56:28

In Hive
hive> select cast(1586318188000 as timestamp);
2020-04-08 11:56:28

hive> select cast(1586318188 as timestamp);
1970-01-19 16:38:38.188

 

Issue Links

• links to: [Github] Pull Request #28605 (GuoPhilipse)