A Method for Live SQL Query Subscription in React Native

Abstract

This thesis proposes integrating the well-established embedded SQLite database into React Native in a reactive and unintrusive manner, allowing developers to take advantage of their existing SQL data-wrangling skillset. The current best practices in client-side global data management in React Native, namely Redux and MobX, enables developers to leverage the unidirectional data flow programming paradigm to simplify data organization without hav- ing to maintain a complex web of callbacks and observers to reflect the latest data in the application UI. However, they leave important aspects of data work, such as filtering, joining and aggregating to the developers, leading to ad-hoc data pipelines that vary from project to project in style and efficiency. Using a central registry for query rerunning combined with run-time schema dependency introspection, SQLite query results could be retrieved and processed as though they were reactive arrays that are automatically kept updated. The minimal API of the library encourages developers to do more work in the declarative logic programming paradigm of SQL thus simplifying application code. The performance in the initial version was acceptable with round-trip latencies low enough to not incur noticeable UI jitters. Moreover, the library was structured so that the core query rerunner is framework- agnostic and could be ported to different UI frameworks and even to backend frameworks in future works.