I’ll admit it – when I first heard of DuckDB, I thought it was just another PostgreSQL or MySQL alternative. But after diving deeper, I’m left with more questions than answers. What exactly is DuckDB, and how does it fit into the world of databases and data wrangling?
One of the biggest sources of confusion for me is how DuckDB is often compared to tools like Polars. I mean, I wouldn’t compare PostgreSQL to Pandas, for example. It seems like apples and oranges. But people are using DuckDB for local data wrangling because of its SQL support, which raises some interesting questions.
Is DuckDB really a database? Or is it more of a dataframe API that just happens to use SQL instead of code? And what does this mean for its potential use cases in ETL/ELT, especially when integrated with tools like dbt?
In my mind, Polars is comparable to Pandas, PySpark, and Daft – not to a tool claiming to be an RDBMS. So, where does DuckDB fit in? Is it a weird beast, as the Reddit title suggests? I’d love to hear from others who have experience with DuckDB and can shed some light on its true nature and use cases.