I guess (correct me if I’m wrong) I could have inserted that into the primary key and not used autoinc, but at least this WAY I kept the AtoInc as a clean easy to use value/index

Firstly – sorry for confusion – was getting my head around ON DUPLICATE KEY for the first time

(I originally thought the AutoInc ID WAS a possible unique key – appreciate your feedback – its good when the penny finally drops)

You have to generate a unique key on insert (and that’s not an AutoInc Field)

My example was on a Link table. Questions – Responses (fixed list) – Sources(people)
The Questions , Responses and Sources all have unique ID’s

BUT I couldn’t see how to make any of THESE unique fields, as the same questions, responses and sources occur more than once in the link table.

BUT – a combination of Question AND Source ID was unique
ie a particular Source has either responded to a Question – in which case update
or has not – in which case insert

my Link table now had a new field to hold a unique response (uni_rsp)
uni_rsp was made UNIQUE and was fed the id of the question and the source
uni_rsp=CONCAT(fk_id_qst_rsp,fk_id_src_rsp)

I guess (correct me if I’m wrong) I could have inserted that into the primary key and not used autoinc

atb

Paul

but at least this was I kept the AtoInc as a clean easy to use value

One thing that’s been puzzling me using ON DUPLICATE KEY
as I mentioned above – still learning and would really appreciate some advice
Using ON DUPLICATE KEY requires you create a record with a unique field value – and I would have thought an AUTOINC ID field would suffice, but (as far as I can tell) it doesn’t work.
Using just a primary key autoinc ON DUPLICATE KEY always inserts a new record rather than updating/inserting
I’ve manage to create a unique value (luckily my data has a unique value combination (CONCAT) each time its created)
Q: Can you make ON DUPLICATE key work with just an Autoincrement Primary Key ID?