Luca Clementi <lclement(at)ucsd(dot)edu> writes:
> It seems that the to_timestamp does not work properly in this case,
> when it comes to parsing the hours.
to_timestamp() is not very robust if the input doesn't exactly match
what it expects for the format string. I'm not sure if that's the
issue here, but have you tried just casting the string to timestamp?
That would use the standard timestamp input converter, which is
pretty flexible.
(BTW, there is definitely 0 hope of recognizing a timezone name that's
written in Chinese characters, unless maybe you fool around with the
timezone-abbreviations configuration file.)
regards, tom lane