Closed dustymc closed 7 months ago
AWG today - Use order 1 for flat/aggregators (concatenate if more than one).
AWG meeting
(Do we need a report of records which don't have at least 1 1?)
Records with no identification_order=1 ID:
Summary:
guid_prefix | count
-------------+-------
UAM:Arc | 111
UAM:Inv | 1
MMNH:Mamm | 5
NMMNH:Paleo | 5
UAM:Herb | 1
MSB:Fish | 10
UMZM:Egg | 2
MSB:Herp | 22
CHAS:Teach | 7
MVZ:Fish | 1
BYU:Edu | 32
CHAS:Herb | 3
MVZ:Bird | 40
MSB:Para | 2
Ping
@msbparasites @campmlc @StefanieBond @jtgiermakowski @barke042 @jandreslopez @adhornsby @sjshirar @ccicero @mkoo @droberts49 @camwebb @wellerjes @Nicole-Ridgwell-NMMNHS
UMZM & MMNH cleaned up.
Sorry to miss the discussion today. So what is needed for these with ID rankings with no "1" value? I thought I could search on all MSB records and use Tools -> Manage Identifications to fix in bulk, but that does not provide a means of converting existing identification order to anything other than "0". Can we add the option of converting a list of records to ID value = some other numeric value?
Taxonomy-flattener/cacher rebuilt per https://github.com/ArctosDB/arctos/issues/7695#issuecomment-2077992235, I've got it running on UAM:Ento without kingdom. Ping me if I need to prioritize anything else, otherwise everything should catch up in a couple weeks.
"should catch up in a couple of weeks" - shouldn't all Arctos users be alerted that any searches involving higher taxonomy could be unreliable for this week (in case any reports have been based on such searches, or loans, etc) and a couple more weeks? Seems like a banner alert or something is in order? ack!
@dustymc MVZ:Bird and MVZ:Fish records in your list fixed. Can you please double check that I didn't miss any, i.e., all MVZ records should have one ID order = 1, and I don't expect that there should be any records with more than one ID order = 1 but please check. Thanks!
@ccicero I think this form is safe-ish for writeSQL:
select
guid,
guid_prefix
from
flat
left outer join identification on flat.collection_object_id=identification.collection_object_id and identification_order=1
where identification.identification_id is null
order by guid
There are no MVZ, you're good, thanks!
Fixed for CHAS records
@dustymc thanks!
Fixed for NMMNH:Paleo records
Completed
UAM:Herb:51396 fixed
From @ccicero
I am currently aggregating "accepted" IDs. From eg
I get...
... so eg
flat.genus
gets the value "Columbarium; Coluzea" - a list of all "accepted" information.I think that it's probably correct to include all the data-bits in all the information-pigeonholes, but I'm not sure that's actually what anyone wants.
Would it be better if I just flattened one "best" identification?
I feel like that's acceptable because anyone who wants the details probably isn't going to get them from the weird "traditional taxonomy" columns anyway, they'll just use the full structured data included with every record (example below).
At this moment, it seems like this would simplify some things (DWC, reports, etc.) without much detriment, and possibly it would even encourage better data management.
Thoughts?
Also there seems to be a bug (I think from the latest DB upgrade) so the data in Arctos are inconsistent. I'll make sure to update anything that needs it when we get this worked out.