Query gridwatch

gridwatch.sql

-- The following queries were entered into https://shell.duckdb.org/ 
-- on an Macbook Pro 2.6 GHz 6-Core Intel Core i7 
-- with a 54Mbps internet connection

select max(wind) 
from 'https://raw.githubusercontent.com/RobinL/iris_parquet/main/gridwatch/gridwatch_2023-01-08.parquet';
-- Takes 6 seconds on the first query, 200ms on subsequent similar queries

select * 
from 'https://raw.githubusercontent.com/RobinL/iris_parquet/main/NSPL/NSPL.parquet' 
where pcd = 'SW1A1AA';
-- Takes 13 seconds on the first query, 100ms on subsequent similar queries

select * 
from parquet_metadata('https://raw.githubusercontent.com/RobinL/iris_parquet/main/NSPL/NSPL.parquet');
-- Doesn't (yet) support display of metadata.  In python can write `pq.read_table("iris_with_metadata.parquet").schema`

view_metadata.py

import pyarrow.parquet as pq
pq.read_table("iris_with_metadata.parquet").schema

# Returns
# sepal_length: double
#   -- field metadata --
#   description: 'The length of the sepal in centimeters'
# sepal_width: double
#   -- field metadata --
#   description: 'The width of the sepal in centimeters'
# petal_length: double
#   -- field metadata --
#   description: 'The length of the petal in centimeters'
# petal_width: double
#   -- field metadata --
#   description: 'The width of the petal in centimeters'
# species: string
#   -- field metadata --
#   description: 'The species of the iris plant'
# -- schema metadata --
# Dataset Description: 'The iris dataset'