Invalid UPCs in the DB seem to follow a consistent pattern

[OptimusEd]

There was an earlier issue 'Raises error if UPC data is greater than 17 characters' and I think it might have helped expose a problem with UPC data in the DB.
Metron-Project/darkseid#206

There seems to be a pattern to invalid UPCs.

Consider these 3 issues:

X-Men (2019) issue 1
https://metron.cloud/issue/x-men-2019-1/
https://www.comics.org/issue/2024862/

For illustrative purposes I've split the UPCs. First 11 digits - product, publisher and series info. 12th digit - check digit. Last 5 - issue info. Should add up to 17 digits.

GCD UPC (17 digits):    75960609466___0_00111
Metron UPC (18 digits): 75960609466_1_0_00111 - extra 1

The whole volume follows this pattern

Avengers (2017) issue 2
https://metron.cloud/issue/avengers-2017-2/
https://www.comics.org/issue/1657231/

GCD UPC (17 digits):    7_59606_08603___0_00211
Metron UPC (18 digits): 7_59606_08603_1_0_00211 - extra 1

King Spawm (2021) issue 49
https://metron.cloud/issue/king-spawn-2021-49/
https://leagueofcomicgeeks.com/comic/4327136/king-spawn-49

LoCG UPC (17 digits):   7_09853_03170_____3_04911
Metron UPC (20 digits): 7_09853_03170_151_3_04911 - extra 151

in all these examples there are extra digits between the first 11 digits and what's supposed to be the 12th digit.
Looking at the these pages' history, I assume they were created (and UPCs were added) automatically. Could it be an issue with whatever automatic tool was used that causes incorrect UPCs to be written?

[bpepple]

The most likely cause is that from whatever source the information was from, be it LoCG, GCD, Diamond, Midtown, etc., the information was incorrect. On pretty much every site, that field a just a free form so it's prone to errors. I've seen distributor codes switch with UPC codes on LoCG, Diamond having more than 20 numbers, and so on.