Batch Edit bug deletes fields when batch-adding a new field

[Tim. Reichard]

I think I have found a nasty bug caused by a Batch Edit operation.  When I batch-edit observations to add a new field, it adds that field to the observations correctly, while erasing other fields previously added to the observations.

I discovered tonight that many (possibly several hundred) of my observations from the current year that previously had the Identification Source field included no longer do.  A few weeks ago, I did a series of batch edits to my Lepidoptera observations from this year to add the Insect Life Stage field.  I gave them all the "adult" value of the field via Batch Edit. Then I changed the few that were actually other stages like caterpillars, not using Batch Edit, but changing them one by one from the individual observation pages.  Many of these observations had the Identification Source filled in beforehand, but now those fields are gone.  They now have only the Insect Life Stage field.

I used the following test to see if field loss was caused by the Batch Editing of my observations.  I found that it is likely the culprit.

I picked 3 observations and removed all fields, one by one (not using Batch Edit). Then I added the Identification Source field, one by one (not using Batch Edit).
I went to my observations page
http://www.inaturalist.org/observations/treichard
and entered some search parameters to put those 3 observations in the first page of the results.
I clicked on Batch Edit, selected those 3 observations, and clicked on Edit Selected.

Next I saw the Batch Edit page.
I opened the Description, Tags, & More box on each observation. Each box showed that the Identification Source field was the only field included with the observation.
In the batch settings at the top, I added the Insect Life Stage field and entered a value of "adult".  I clicked on Apply Fields.
I looked again at the Description, Tags, & More boxes, and all now showed both fields for each observation. So far so good.
I clicked Save All.

Next I looked at the individual observation pages to see what happened.
On each, below the photos at the upper left, the Insect Life Stage field appeared, and the Identification Source field was gone.

I repeated this test with other observations and the field loss was the same. So the bug is reproducible.  I also looked at non-Lepidoptera observations that I did not apply a batch edit to, and they still have the Identification Source field.

So I think this is how I lost the Identification Source fields from many of my 2016 Lepidoptera observations.  If it is possible to have them restored from a backup, without reverting other observation details, I would greatly appreciate that, or to know that it can't/won't be done.

Tim

[Tony Iwane]

Hi Tim,

I was able to reproduce your bug and I notified our team - so sorry that happened. We'll look into the issue and let you know.

Tony

[Charlie Hohn]

i just lost my fields from 200 observations because of this bug - some of which is data I will never get back. Really sad that this nasty bug has been here for almost a year now, and surprised I didn't run into it sooner. PLEASE fix this.

[Charlie Hohn]

ok, i have to apologize and admit confusion. it looks like my lost fields came back after a while! I don't understand at all but am glad it happened

[Charlie Hohn]

This still seems to be happening, i think?. I added 'secondary natural community' field to a batch of obs and that deleted my other field 'natural community' from them. In this case easy to fix, but this bug can cause a lot of lost data

[tony rebelo]

I also lost about 50 observation fields yesterday, but i dont know what I did.   

I entered them at 3pm (one by one) via a project, but when I summarized them at 6pm none of them existed.  (at about this time there was a change in how the observations fields displayed: instead of being confined to the Projects, they started showing up under the Observation Fields again - they had not all week).

I never touched these fields in the interim, although I did edit the project. 

I have since edited them and they are behaving themselves.

re: https://www.inaturalist.org/observation_fields/7982

[Tony Iwane]

Very sorry about this, folks, this issue has been on our radar but got lost in the shuffle. I've bumped it back up to the dev team.

Tony Iwane

[Charlie Hohn]

Thanks! I do use the fields a lot, especially for classifying natural community/habitat type

[Tony Iwane]

Alright, we should have it fixed. I tested it in Staging and on the website and it's looking good. Maybe try testing it out with a few observations at first and let us know if you encounter any bugs.

Tony

[Chris Vynbos]

Just (22 July 2018) added a tag via a batch edit and the existing tags vanished. Did I do something wrong? 

[Tony Iwane]

Hey Chris,

I can replicate that, looks like a bug. I'll file a bug report for it, sorry about that.

Tony Iwane

[Chris Vynbos]

Thanks Tony I, But if you solve that bug, what happens if you want to delete an incorrect tag off multiple obs? If batch editing doesn't delete existing tags, how does one delete an incorrect tag (maybe misspelled) across multiple obs? One needs the batch edit tag function to have an 'add' mode and a 'delete' mode where the latter will delete all instances of whatever tag is typed in, while the add mode will add them. 

[Tony Iwane]

Hi Chris,

Well, the deleting tags when batch adding tags has been fixed but yes, this does mean you can't batch delete tags currently. I think it's better to err on the side of not deleting, with the way the current system is set up so that no one accidentally deletes a bunch of tags in the future. 

Tony Iwane

[Chris Vynbos]

Just to mention that because one can't batch delete tags I'm going to have to manually delete 40 incorrect tags , so my request still stands "One needs the batch edit tag function to have an 'add' mode and a 'delete' mode where the latter will delete all instances of whatever tag is typed in, while the add mode will add them." Luckily 40 isn't too many, but it is a hassle and could have been many more.