Bridging the Gap between SQL and R
summary
Introduces queryparser, a pure‑R, no‑dependency SQL parser that converts SELECT clauses into unevaluated R expressions, and tidyquery, which uses dplyr and rlang to execute SQL directly on R data frames without copying them to a DB. Parsing time is roughly 100–300 ms for typical queries. Generates runnable dplyr code (show_query) so SQL can be translated into idiomatic dplyr, handles joins and aggregates, and preserves lazy backends like data.table and Arrow; some SQL features and subqueries remain unimplemented.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 0:49Show of Hands
- 2:11SQL vs R
- 5:35DBplier
- 7:17SQLDF
- 10:32Query Parser
- 12:08How Parser Works
- 16:33TidyQuery
related talks
18mduckplyr: Tight Integration of duckdb with R and the tidyverse
14mAutomated Translation of Functional Big Data Queries to SQL
06mBuilding a SQL editor around “fast”
32mBecause SQL is everywhere but you don't want to use it
32mPipes or Brackets: Dplyr and Data.Table
49mSQLPage — Building a full web application with nothing but Postgres and SQL queries
48mCompositional, Expressive, Performant Database Queries: Choose 3
19mIn-Process Analytical Data Management with DuckDB