Mo Sarwat
Parquet is a columnar format for storing data in an object store in the cloud. It’s definitely not as fast as traditional OLAP databases, but again real-time queries is not the purpose of parquet in the first place. DuckDB is another story. It is an OLAP query engine that runs in your application on small data. It can also pull data from small-scale parquet files - a couple of hundreds of GBs maybe. After that, it cannot easily scale. Works well on the client side, but not suitable for scale-as-you go cloud applications. But, the combination of DuckDB+Parquet are definitely a good fit for the examples that you showcase sometime. So, kudos on using both of them and keep it up.