[DRILL-1216] Decimal multiplication breaks when the precision exceeds 38 even when the actual data is small

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.5.0
• Component/s: Execution - Data Types
• Assignee: Unassigned
• Reporter: Rahul Kumar Challapalli
• Created: 29/Jul/14 17:45
• Updated: 07/Aug/14 03:59
• Resolved: 07/Aug/14 03:59

Description

git.commit.id.abbrev=caa8b78

Input File Contents :

111.11|222.22|333.33
444.44|555.55|666.66
777.77|888.88|999.99

The below SQL query gives inaccurate results since the precision of the multiplication is 39 (>38):

select cast(columns[0] as decimal(15,2)) * cast(columns[1] as decimal(15,2)) * cast(columns[2] as decimal(15,2)) val from `decimal_scale.tbl`;
+------------+
|     val      |
+------------+
| 1000000000 |
| 0          |
| 0          |
+------------+

However the below SQL query succeeds :

 select cast(columns[0] as decimal(5,2)) * cast(columns[1] as decimal(5,2)) * cast(columns[2] as decimal(5,2)) val from `decimal_scale.tbl`;
+------------+
|     val      |
+------------+
| 8230205.763786 |
| 164604115.275720 |
| 691337284.158024 |
+------------+