https://bugs.documentfoundation.org/show_bug.cgi?id=136615
--- Comment #46 from Eike Rathke <er...@redhat.com> --- (In reply to Albrecht Müller from comment #45) > Just to be sure: If the SECOND function is implemented to always truncate as > you proposed in comment #39 then this relation would also hold for the > SECOND function. Right? Right. > I want to understand this clearly because I think > that this half second is related to your comment #35. ? I don't think so. My comment 35 is about floating point being an inadequate data type for date+time, as many if not most values (even in the seconds range), or rather all values that are not multiples of denominators that are a power of 2 (x/2, x/4, x/8, ...), can not be represented as distinct accurate binary floating point values, hence calculating with these values is even less accurate and accumulates round-off errors as with any floating-point calculation. Any final rounding is just an inaccurate workaround to calm the user and pretend that results would be accurate numbers, which they aren't. Read material linked under https://erack.de/bookmarks/D.html#010203 The relation of the half second you mentioned is simply truncate(x+0.5) == round(x) -- You are receiving this mail because: You are the assignee for the bug.