I'm just really happy about the allowing of update when inserts conflict. This will tidy a lot of code up for me and make writing in the future that much neater.
Anyone have ever built an application utilising row level security?
In old-school systems you normally have one database user and a connection pool set up in the app with that user. Row level access is handeld with join tables to sort out what data is accessible for the specific users (sent as parameter to most queries).
With row-level security you would crate a db-user for each application user and also its own connection pool?
Sure, but this is about the new release — I don't see the point of having two threads about the same thing. (The press release links to the release notes, too.)
Personally, I always read the release notes but am less interested in the press release. I think many others prefer the opposite, hence my posting both. You raise a valid point though.
I'm excited about the sorting improvements. Large index creation performance has been an embarrassment (I often have to defend pg against Oracle!), hopefully this will make things vastly better.
10 comments
[ 3.2 ms ] story [ 28.1 ms ] threadhttp://www.postgresql.org/docs/9.5/static/sql-insert.html
In old-school systems you normally have one database user and a connection pool set up in the app with that user. Row level access is handeld with join tables to sort out what data is accessible for the specific users (sent as parameter to most queries).
With row-level security you would crate a db-user for each application user and also its own connection pool?
This allows pooling.
I made a quick app with pg 9.5 rc1, rls, spring data rest which I'm hoping to share next week.
Personally, I always read the release notes but am less interested in the press release. I think many others prefer the opposite, hence my posting both. You raise a valid point though.