You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This is a JIRA to discuss the behavior of operator kernels that require more than one decimal column input where the column types have a different scale parameter.
For example:
a: decimal(12, 2)
b: decimal(10, 3)
c = a + b
Arithmetic is the primary use case, but anything that needs to efficiently operate on decimal columns with different scales would require this functionality.
I imagine that @jacques-n and folks at Dremio have thought about and solved the problem in Java. If so, we should consider implementing this behavior in C++. Otherwise, I'll do a bit of reading and digging to see how existing systems efficiently handle this problem.
These are the notes that we have about our scale/precision behaviors (we're actually in the middle of reimplementing decimal using llvm). These are primarily modeled after MS SQL Server's behavior. We can share as we've reimplemented.
This is a JIRA to discuss the behavior of operator kernels that require more than one decimal column input where the column types have a different
scale
parameter.For example:
Arithmetic is the primary use case, but anything that needs to efficiently operate on decimal columns with different scales would require this functionality.
I imagine that @jacques-n and folks at Dremio have thought about and solved the problem in Java. If so, we should consider implementing this behavior in C++. Otherwise, I'll do a bit of reading and digging to see how existing systems efficiently handle this problem.
Reporter: Phillip Cloud / @cpcloud
Related issues:
Note: This issue was originally created as ARROW-1761. Please see the migration documentation for further details.
The text was updated successfully, but these errors were encountered: