DM-15760: Change first/last validity from exposure ID to datetime #90
Conversation
timj
changed the title
config/schema.yaml
Outdated
| First exposure identifier included in the range (inclusive). may | ||
| be zero to indicate an open interval. | ||
| Date of first exposure included in the range (inclusive). May | ||
| be far in the past to indicate an open interval. |
There was a problem hiding this comment.
Should we say TAI explicitly here to avoid any possibility of confusion in the future?
And this is mostly off-topic, but it got me thinking: do you think we should by using MJDs for all of our datetimes in the Registry to avoid the possibility of people using database datetime UDFs that are inappropriate for our time systems?
There was a problem hiding this comment.
TAI MJD might be safer (and we don't need millisecond accuracy). Shall I make that change on this ticket?
There was a problem hiding this comment.
You're welcome to make it on this ticket as far as I'm concerned, but note that that will involve some code changes, maybe including some that extend beyond daf_butler, as we actually are populating the datetime_begin (etc) fields.
Just curious: what kind of accuracy would that give us?
There was a problem hiding this comment.
Single 64bit float using MJD is about 1 micro second accuracy I think. Integer MJD day and 64-bit float for fractional day seems to be 0.02 ns.
timj
force-pushed
the
tickets/DM-15760
branch
from
151a467
to
f65cac8
Compare
No description provided.