Thread

Re: Numeric Type Precision Not Respected in Function or Procedure Arguments

David G. Johnston <david.g.johnston@gmail.com> — 2025-09-22T21:33:45Z

On Monday, September 22, 2025, Aaron Ackerman <aackerman@goodmorning.com>
wrote:

> When a NUMERIC type is used as a function or procedure argument, the
> value does not actually follow the user-defined precision, such as
> NUMERIC(10,4).
>

This is a known limitation - the “typmod” (type modifier) is not stored in
the metadata for a function’s interface.

Pretty sure it’s documented but not able to go hunting for it at the moment.

David J.