Posts Tagged "optimization fence"

PostgreSQL’s CTEs are optimisation fences

As part of AXLE project work I recently found myself re-examining issues around optimisation fencing and non-semantic query execution changes in PostgreSQL. One key issue was the inability to use a CTE where optimisation (inlining, push-up/pull-down, flattening, etc) was desired. I’d like to explain that in more detail, as it’s something that can surprise new…