Paused due to POC of Elastic Search. Bugfest: Inventory. Effective Location Result Counts Don't Add Up (Off by Unacceptable Amount) And Seem Suspicious
Description
CSP Request Details
CSP Rejection Details
Potential Workaround
Attachments
is blocked by
relates to
Checklist
hideTestRail: Results
Activity

Charlotte Whitt May 18, 2022 at 1:50 PM
Bug closed while the problem was solved when we swap'ed to Elasticsearch.

Cate Boerema August 17, 2020 at 8:03 AM
There were issues related to counting.
Adam Dickmeiss can you please help to identify if we have a related issue open?Then we can make a ticket for investigation.
and , I'm just following up on this. Is Core Platform planning to take a look at this and when?

Oleksii Popov August 5, 2020 at 1:16 PM
There were issues related to counting.
can you please help to identify if we have a related issue open?
Then we can make a ticket for investigation.

Craig McNally August 5, 2020 at 12:23 PM
I'm not sure - it feels like vacations have had a fairly significant impact our team's capacity lately. and would probably be better equipped to provide an answer.

Cate Boerema August 4, 2020 at 11:17 AM
Excellent. Thanks for your analysis, and for assigning this to Core Platform.
, any chance these could be looked at in time for the Q3 platform deadline?
Thanks!
Details
Details
Assignee
Reporter

Tester Assignee

Steps to reproduce:
Login to FOLIO BugFest, go to Inventory
Select the Item segment
Test filtering for "Effective location (item)"
Select multiple locations
Try:
UC/HP/ASR/ACASA
UC/HP/ASR/ASRHP
Notice the total result count is "5,593,722 records found"
Now filter by each separately and notice the resulting counts:
UC/HP/ASR/ACASA - 8,338 records found
UC/HP/ASR/ASRHP - 5 records found
Manually added together, this is 8,343 records (a very far cry from 5,593,722)
Now try two totally different locations:
UC/HP/ASR/Atk
UC/HP/ASR/JRLASR
Notice the total results count is exactly the same as the total for the other two locations ("5,593,722 records found"). Even knowing result counts are estimates, this seems really odd.
Now filter by each separately and notice the resulting counts:
UC/HP/ASR/Atk - 8,338 records found This is exactly the same result as UC/HP/ASR/ACASA which is unlikely.
UC/HP/ASR/JRLASR - 92,998 records found
Manually added together, this is 109,682 (also very far from 5,593,722)
Expected Result:
Then the filter box will display the selected locations, and the search result will be the combined search (boolean operator: OR), so that means Location X OR Location Y.
Even knowing the result counts are estimates, the combined totals should be MUCH closer than the reported total than they are
It is highly suspect that result counts for different location selections are exactly the same. It would be great if we could look into the DB to see how they compare to reality.
Actual Result:
See repro steps above
Additional Info:
This may have the same underlying cause as Bugfest: Inventory. Item segment. Filtering on Suppress from discovery: Yes and No Don't Add Up
The issue can not be reproduced in FOLIO Snapshot, but in uChicago's sandbox environment (Edelweiss) I see the same bug - except here I get 6,369,032 records found.
Screenshots:
BugFest:
https://bugfest.folio.ebsco.com/inventory?filters=effectiveLocation.d695b3e4-5bbb-4ab2-b998-6e52cd395d86%2CeffectiveLocation.e5d578f4-17ce-4c70-b1b5-565f3605e10b&sort=Title
https://bugfest.folio.ebsco.com/inventory?filters=effectiveLocation.f369266a-a209-4e4a-b487-d1acf3ee6857%2CeffectiveLocation.c3dd9997-463b-47e3-958c-2c6fc2775f90&sort=Title
uChicago: