[bug #61698] Mean operator not returning NaN for blank regions when input is integer

[Mohammad Akhlaghi]

URL:
  <https://savannah.gnu.org/bugs/?61698>

                 Summary: Mean operator not returning NaN for blank regions
when input is integer
                 Project: GNU Astronomy Utilities
            Submitted by: makhlaghi
            Submitted on: Sat 18 Dec 2021 05:36:56 PM UTC
                Category: Arithmetic
                Severity: 3 - Normal
              Item Group: Output not reasonable
                  Status: In Progress
                 Privacy: Public
             Assigned to: makhlaghi
             Open/Closed: Open
         Discussion Lock: Any

    _______________________________________________________

Details:

When the inputs to the 'mean' operator don't contain any data in a pixel, the
output pixel should have a NaN value (the output of mean is always a floating
point numerical data type).

However, when the inputs are integer types, this doesn't happen and large
negative values are used instead (for example -2.14748e+09 when the inputs are
32-bit integers). 

This bug was reported by Zohre Ghaffari.




    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?61698>

_______________________________________________
  Message sent via Savannah
  https://savannah.gnu.org/