@hawksthree made an identification on one of my observations (https://www.inaturalist.org/observations/69211703), and noted that the order of the identifications was odd, since my identification (which should be first) shows up second.
I made my identification as part of uploading, and the timestamp matches the observation submit time. hawksthree’s identification has a timestamp that is somehow earlier than the observation submit time.
hawksthree’s comment about the identification order strangeness seems to have a reasonable timestamp, but then my reply comment has a timestamp that is prior to the observation submit date.
the observation submit date is 2021-02-07 11:56:51 (-06:00).
here’s what the identifications / comments look like with timestamps and IDs, in order as they appear on the observation detail page:
||Item + Sequence
||2021-02-07 11:28:10.628 (-06:00)
||strange. thanks for pointing that out and for the ID.
||2021-02-07 10:41:24 (-07:00
||(species) Spatula clypeata / Northern Shoveler (supporting)
||2021-02-07 10:56:51 (-07:00)
||(species) Spatula clypeata / Northern Shoveler (improving)
||2021-02-07 12:00:58.279 (-06:00)
||that’s weird - I just agreed with your call on N.S. and it switched the ID sequence (?)
additional examples (my observations):
additional examples (observed by others):
looking at a sample of identifications within a range of IDs that occurred during the same time period as the examples above, i would expect the identifications to be effectively sorted by datetime when the IDs are sorted, but that is not the case. there are occasional cases when an identification’s timestamp appears to be off by about 18 minutes vs its ID neighbors. see: https://jumear.github.io/stirfry/iNatAPIv1_identifications.html?id_above=153827846&id_below=153828406&order_by=id.
a similar problem also seems to extend to observation submit dates. for example, it looks like a user loaded 42 observations at the same time, but a few of them have a different submit time (and update date) recorded: https://jumear.github.io/stirfry/iNatAPIv1_observations.html?user_id=yellowrattle&id_above=69210529&id_below=69213081&order_by=id&per_page=50. i was thinking the dates should be off by about 18 minutes, just as with the identifications that were created around the same time, but they’re off by more than that (double?) in this case.
Maybe the person’s time or your time is wrong?
i doubt that the timestamp on these kinds of things is based on the time from each person’s device. it would make more sense for that to come from servers.
if you look at my main example, you’ll see that the time is right for my identification but not for my comment, and vice versa for the other user. also, it’s not every observation that has this problem, even though the same user identified multiple observations of mine from the same set. so if the time were coming from a local device, i would expect the times recorded to be consistently wrong or consistently right, and that’s not the case.
it seems more likely that either multiple servers / processes are picking up the requests and their clocks are out of sync, or maybe the code is doing something strange when it records the timestamp.
i don’t have a good way to query for cases where the identification datetime is earlier than the observation create date, but from my limited sampling, it looks like the issue seemed to happen between about 11:30 AM and 12:30 PM CST on 2021-02-07. i couldn’t find examples at other times a few hours before or after that time period, and i couldn’t find any examples during that time period on the previous 2 days either.
Added to my weekly report.
i see kind a similar problem. year of the dates repeating itself i think.
this is turkish language version
this is english language version
another issue is when suggested ID is longer than line, it gets kind a hidden in turkish language.
I got an incorrect observation date and I cannot fix it in Edit mode. I am using the web UI with the latest (Chromium-based) version of Microsoft Edge browser. I open the observation, click Edit and enter the corrected date and time as in the example below that field, Save the observation, but the Observed date and time do not change. Clicking Edit again shows my corrected date and time, but it does not show on the regular Observation view - it still shows the old data.
- While I am signed in to iNaturalist, I go to https://www.inaturalist.org/observations/69629216
- Date and time shown are Feb 12, 2021 · 10:11 AM EST, which is incorrect
- Click Edit
- Enter the correct date and time in the suggested format: 2021-02-13.
- Click Save
- Observation date and time are unchanged
- Click Edit again
- Correct date and time shown on this screen but not after clicking Save
Note: Deleting all browser history and cache did not help, and neither did signing out of iNaturalist and signing back in again .
the thing you’re reporting sounds like a translation issue, unrelated to what i originally reported. you may want to log your issue separately.
this also sounds like an issue unrelated to the original reported issue.
you may want to log this separately. you might want to try changing the date/time on another observation to see if it’s isolated to only one observation or if the problem occurs for any observation. based on what you’re describing, i’m guessing you’re probably inputting an invalid datetime.
looking at https://api.inaturalist.org/v1/observations?id=69629216, it looks like you’re inputting an invalid datetime:
"observed_on_string": "2021-02-13:13:50:45". (replace the
: between the date and the time with a space.)
my issues are resolved and yes it was mistranslation.
Have you noticed this happening any more? We’re thinking it probably was a temporary issue with a server clock, according the the logs.
i haven’t noticed any other cases, but i haven’t really been looking. you all would probably be in a better position to see if it is still happening, since you can probably query the database for observations with identifications timestamped before the observation create datetime.
sounds reasonable, but the thing that doesn’t really make sense to me is why only some of the observations, etc. created during the period were affected…