Used a natural PK unique_id instead of id
As discussed on Sql WG on glip
Is more performant to use a natural PK like unique_id instead of id
cause we need to know if an unique_id already exits to decide if insert or update the #__jstats table
so is better to use unique_id as a PK instead of adding another key
ps
The perfomance may vary a lot depending on server settings, so if we are concerned about performance
I think this is a good use case to use Redis instead is superfast
Used a natural PK
unique_id
instead ofid
As discussed on Sql WG on glip Is more performant to use a natural PK likeunique_id
instead ofid
cause we need to know if anunique_id
already exits to decide if insert or update the#__jstats
table so is better to useunique_id
as a PK instead of adding another keyps The perfomance may vary a lot depending on server settings, so if we are concerned about performance I think this is a good use case to use Redis instead is superfast