https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=5670 --- Comment #71 from Alex Sassmannshausen <alex.sassmannshausen@ptfs-europe.com> --- (In reply to Jonathan Druart from comment #67)

(In reply to Alex Sassmannshausen from comment #66)

...

(In reply to Jonathan Druart from comment #65)

...

(In reply to Jonathan Druart from comment #64)

...

...

My alternative proposal now would be: - create a new custom table "housebound_role" - with three columns (borrower_id, deliverer, chooser)

The PK for this new table would be borrower_id, so should not it be 2 new columns in the borrowers table instead?

I considered that, but proposed a new table because: - the borrowers table is quite large as it is. Adding columns to it would make it slightly less discoverable - the vast majority of borrowers in the system will have neither a "deliverer" nor a "chooser" role, so those columns will just be empty/false in the borrowers table (this is particularly the case if the housebound module is switched off!). - in the housebound_role table the columns will only exist if the borrower is in fact a volunteer of some form (either chooser or deliverer), so each entry will be "doing something".

Conversely, joins are quite cheap, so I think I prefer this design.

My concern is more about the DB normalization, I think we need to keep that in mind and normalize it as much as possible to keep data integrity. I'd like to get more opinions before you start.

Fair enough, I've opened a thread on Koha-Dev. Let's see where that goes. If worst comes to worst we might be able to hammer it out at the hackfest. -- You are receiving this mail because: You are watching all bug changes. You are the QA Contact for the bug.