=?UTF-8?Q?R=C3=A9mi_Cura?= <remi.cura@gmail.com> writes: > I tried to reproduce, and as you pointed out, updating the function in > place is not what dropped the columns. > Instead, I must have dropped and recreated the function, which indeed drops > the generated columns. > I still feel this is a very dangerous behavior. Yeah, we changed this at v15. See https://www.postgresql.org/message-id/flat/17439-7df4421197e928f0%40postgresql.org https://git.postgresql.org/gitweb/?p=postgresql.git&a=commitdiff&h=cb02fcb4c regards, tom lane
pgsql-bugs by date:
Соглашаюсь с условиями обработки персональных данных