Skip to main content

Query files

DuckDB support querying multiple file formats. This example show how to use it to query parquet files.

Apache parquet is an efficient file format to store column-oriented data.

Download Parquet files​

Download Yellow Taxi record data

wget https://d37ci6vzurychx.cloudfront.net/trip-data/yellow_tripdata_2024-01.parquet

Run µQuery with files​

docker run -p 8080:8080 -v ./yellow_tripdata_2024-01.parquet:/tmp/yellow_tripdata_2024-01.parquet fb64/uquery

Query parquet files​

curl --location 'http://localhost:8080' \
--header 'Accept: application/json' \
--header 'Content-Type: application/json' \
--data '{
    "query":"select * from read_parquet('\''/tmp/yellow_tripdata_2024-01.parquet'\'') limit 10"
}'

Note that DuckDB HTTPFS extension is preinstalled on docker image, so you can directly query file over http(s)

select * from read_parquet('https://d37ci6vzurychx.cloudfront.net/trip-data/yellow_tripdata_2024-01.parquet') limit 10