Recover previous Codex session

Execute queries on Arrow Parquet files without full loading

Fast analytics on columnar Parquet data without loading entire files into memory

• · Parquet file path(s)
• · SQL query
• · Optional filters

• · Query result set
• · Partial file scan output

• · Apache Arrow
• · SQL execution engine
• · Parquet reader

• · Parquet files on local/remote filesystem
• · SQL query syntax valid

• · Query references missing column
• · File schema incompatible
• · Memory exhausted on large result

• · Zero-copy Parquet access
• · Push-down filtering support
• · CLI and library APIs available