Parquet Data Analysis with DuckDB WASM

Analyze remote Parquet files using SQL queries with DuckDB WASM.

Enter the URL of a remote Parquet file to analyze
Use {SOURCE} as placeholder for the Parquet URL
Initializing DuckDB...
Common SQL Functions:
  • SELECT * FROM 'url' - Read data directly from Parquet URL
  • COUNT(*) - Count total number of records
  • DESCRIBE SELECT * FROM 'url' - Show column information
  • WHERE column = value - Filter records
  • ORDER BY column DESC - Sort results
  • GROUP BY column - Group data for aggregation
  • LIMIT n - Limit number of results
Partitioned Data Benefits:
  • Partition Pruning: Query only specific models/grids with WHERE model_name = 'WRF-NARR_HIS'
  • Cross-Model Analysis: Compare models easily with GROUP BY model_name
  • Efficient Filtering: DuckDB automatically skips irrelevant files when filtering by partition columns
  • Standardized Schema: All files have consistent model_name, grid_name metadata columns
  • URL Pattern: Use URL encoding: model_name%3DWRF-NARR_HIS for model_name=WRF-NARR_HIS
Query Results