So, I'm working on using Jooq to create a caching layer over Postgres. I've been using the MockConnection/MockDataProvider objects to intercept every query, and this is working, but I'm having a few issues.
First, how do I determine between reads and writes? That is, how do I tell whether a query is an insert/update/etc or a select, given only the MockExecuteContext that's passed into the execute method in MockDataProvider?
And I'm a bit confused on how I can do invalidations. The basic scheme I'm implementing right now is that whenever a "write" query is made to a table, I invalidate all cached queries that involve that table. This goes back to my first question, on telling different types of queries from each other, but also brings up another issue: how would I identify the tables used in a query given only the sql string and the bindings (both are attributes of MockExecuteContext)?
Also, is this a correct approach at caching? My first thought was to override the fetch() method, but that method is final, and I'd rather not change something already embedded in Jooq itself. This is the only other way I could think of to intercept all requests made so I could create a separate, persistent caching layer.
I have seen this (https://groups.google.com/forum/#!topic/jooq-user/xSjrvnmcDHw) question, but I'm still not clear on how Lukas recommended to identify tables from the object. I can try to implement a Postgres NOTIFY, but I wanted something native in Jooq first. I've seen this issue (https://github.com/jOOQ/jOOQ/issues/2665) pop up a lot too, but I'm not sure how it applies.
Keep in mind that I'm new to Jooq, so it's quite possible that I'm missing something obvious.
Thanks!