Speaker Interview with László Forró
- Caching long running queries
- Wednesday 14:40
- Could you briefly introduce yourself?
- I am currently a back end engineer, involved in our data modelling and database with PostgreSQL, also our back end code. My interest also extends to other fields like language processing and information retrieval, also lately with ML/AI.
- How do you engage with the PostgreSQL Community?
- Through mailing list, conference.
- Have you enjoyed previous pgconf.eu or FOSDEM conferences, either as an attendee or as speaker?
- Yes, I have visited pgcon.eu conferences as attendee and I had a ligtning talk last year in Lissabon on a simple algorithm for partitioning live tables on the fly and published with my colleague.
- What will your talk be about, exactly? Why this topic?
I would like to give a short overview of PostgreSQL related caching strategies and place our solution that we implemented as extension for caching queries in this picture.
Our problem is that we have expensive queries returning result sets, where the entities are without primary keys, so a number of possibilities are excluded.
On the top of it our queries are dynamically mapped between back end objects so we can have 10k queries of similar structure but addressing various data sets.
And unfortunately, we have a sporadic use pattern, so no ‘hot’ data that would fit easily into the caching scenarios we have so far investigated.
- What is the audience for your talk?
- Database and back end developers.
- What existing knowledge should the attendee have?
- No very specific knowledge as a developer is required beyond that is usual at the PgConfs.
- What is the one feature in PostgreSQL 12 which you like most?
- Inlined CTEs, as the developer must know less _how_ the particular query syntax is executed, only what the query should do. But PostgreSQL has a number of improvements each year that are remarkable.
- Which other talk at this year's conference would you like to see?
- Apart from database administration issues - that I am not much involved in - I am very interested in various talks on the list. Not yet decided what I can visit.
- Which measure, action, feature or activity would—in your eyes—help to accelerate the adoption of PostgreSQL?
Calling the developer’s attention outside the database world that PostgreSQL is more than a place where the data is stored in tables and retrieved using that strange SQL language. I feel a cognitive gap between those who are experienced with software architecture and development but not involved in the database, data modelling and the wide range of possibilities. A good database design and features that PostgreSQL has, can reduce the amount of code between the front end and the database. PostgreSQL has a lot more to offer than a fancy persistence layer but the system architects and designers must also be aware of it.
As features... IoT gains importance, characterized by high insert rate of normally small amount of data. The other is data analysis, perhaps on the data from IoT devices. Perhaps some discussion if the latter has place in PostgreSQL, and if yes, to what extent, would be interesting.