Building a JSONiq Query Optimizer using MLIR

Semi-structured data formats like JSON gained popularity through their ability to represent arbitrarily complex data in a way that it can easily be read and written by humans, and parsed and generated by machines. This simplicity is especially useful for applications where it is not worth to spend time in schema design and data migration. However, it comes at a price: Query execution is much slower. In this bachelor's thesis we apply some optimizations on a MLIR dialect for JSONiq. We also take a closer look at type inference for a selection of JSONiq expressions