feat: add JSON lines, Arrow IPC stream, and Parquet output formats

Cargo.toml

@@ -20,6 +20,7 @@ arrow-schema = "59.3.0"
 clap = { version = "4.6.7", features = ["derive"] }
 comfy-table = "8.0.1"
 nu-ansi-term = "0.50.3"
+parquet = { version = "59.3.0", default-features = false, features = ["arrow", "snap"] }
 reedline = "0.52.0"
 syntect = "5.3.0"
 terminal-colorsaurus = "1.0.3"

README.md

@@ -21,7 +21,7 @@ A command-line tool for querying databases via [ADBC](https://arrow.apache.org/a
 - **Interactive SQL shell** - Execute SQL queries with command history and intuitive navigation
 - **Syntax highlighting** - SQL queries highlighted for improved readability
 - **Formatted output** - Results displayed in clean, aligned tables with dynamic column width
-- **File export** - Export query results to JSON, CSV, or Arrow IPC files
+- **File export** - Export query results to JSON, JSON lines, CSV, Arrow IPC, or Parquet files
 - **Fast and lightweight** - Built in Rust for high performance and minimal resource usage
 
 ## Installation
@@ -107,8 +107,11 @@ Execute a query and output the result to a file:
 
 ```sh
 databow --driver duckdb --query "SELECT 42 AS the_answer" --output result.json
+databow --driver duckdb --query "SELECT 42 AS the_answer" --output result.jsonl
 databow --driver duckdb --query "SELECT 42 AS the_answer" --output result.csv
 databow --driver duckdb --query "SELECT 42 AS the_answer" --output result.arrow
+databow --driver duckdb --query "SELECT 42 AS the_answer" --output result.arrows
+databow --driver duckdb --query "SELECT 42 AS the_answer" --output result.parquet
 ```
 
 ## Reference

docs/index.md

@@ -19,5 +19,5 @@ databow is a command-line tool for querying databases.
 - **Interactive SQL shell** - Execute SQL queries with command history and intuitive navigation
 - **Syntax highlighting** - SQL queries highlighted for improved readability
 - **Formatted output** - Results displayed in clean, aligned tables with dynamic column width
-- **File export** - Export query results to JSON, CSV, or Arrow IPC files
+- **File export** - Export query results to JSON, JSON lines, CSV, Arrow IPC, or Parquet files
 - **Fast and lightweight** - Built in Rust for high performance and minimal resource usage

docs/reference.md

@@ -108,11 +108,14 @@ databow --driver duckdb --query "SELECT 42 AS the_answer" --output result.json
 
 The output format is inferred from the file extension:
 
-| Extension       | Format    |
-|-----------------|-----------|
-| `.json`         | JSON      |
-| `.csv`          | CSV       |
-| `.arrow`, `.ipc`| Arrow IPC |
+| Extension       | Format           |
+|-----------------|------------------|
+| `.json`         | JSON             |
+| `.jsonl`        | JSON lines       |
+| `.csv`          | CSV              |
+| `.arrow`, `.ipc`| Arrow IPC file   |
+| `.arrows`       | Arrow IPC stream |
+| `.parquet`      | Parquet          |
 
 ## --help
 

docs/tutorial.md

@@ -143,7 +143,7 @@ $ databow --profile warehouse --file query.sql
 └───────────────────┘
 ```
 
-Instead of printing query results to stdout, the [`--output` argument](/reference/#-output) can be used to write results to JSON, CSV, or Arrow IPC files:
+Instead of printing query results to stdout, the [`--output` argument](/reference/#-output) can be used to write results to JSON, JSON lines, CSV, Arrow IPC file, Arrow IPC stream, or Parquet files:
 
 ```console
 $ databow --profile warehouse --query "SELECT * FROM penguins" --output penguins.csv

src/cli.rs

@@ -1,6 +1,7 @@
 // Copyright 2026 Columnar Technologies Inc.
 // SPDX-License-Identifier: Apache-2.0
 
+use crate::output::OutputFormat;
 use crate::table::TableMode;
 use clap::{Arg, ArgAction, Command, value_parser};
 use std::path::PathBuf;
@@ -84,7 +85,8 @@ pub fn parse_args() -> AppConfig {
         Arg::new("output")
             .long("output")
             .help("Write result to file")
-            .value_name("file"),
+            .value_name("file")
+            .value_parser(parse_output_path),
     ];
     let command = Command::new("databow")
         .version(env!("CARGO_PKG_VERSION"))
@@ -163,7 +165,7 @@ pub fn parse_args() -> AppConfig {
         .copied()
         .unwrap_or_default();
 
-    let output_path = matches.get_one::<String>("output").map(PathBuf::from);
+    let output_path = matches.get_one::<PathBuf>("output").cloned();
     if output_path.is_some() && matches!(query_source, QuerySource::Interactive) {
         eprintln!("Error: --output cannot be used in interactive mode");
         exit(1);
@@ -177,6 +179,12 @@ pub fn parse_args() -> AppConfig {
     }
 }
 
+fn parse_output_path(s: &str) -> Result<PathBuf, String> {
+    let path = PathBuf::from(s);
+    OutputFormat::from_path(&path)?;
+    Ok(path)
+}
+
 fn uri_has_driver_scheme(uri: &str) -> bool {
     let Some(idx) = uri.find(':') else {
         return false;
@@ -202,6 +210,26 @@ fn parse_option(option: &str) -> Result<(String, String), String> {
 mod tests {
     use super::*;
 
+    #[test]
+    fn test_parse_output_path_valid() {
+        assert_eq!(
+            parse_output_path("out.parquet").unwrap(),
+            PathBuf::from("out.parquet")
+        );
+    }
+
+    #[test]
+    fn test_parse_output_path_unsupported_extension() {
+        let err = parse_output_path("out.xyz").unwrap_err();
+        assert!(err.contains("Unsupported file extension"));
+    }
+
+    #[test]
+    fn test_parse_output_path_no_extension() {
+        let err = parse_output_path("out").unwrap_err();
+        assert!(err.contains("no file extension"));
+    }
+
     #[test]
     fn test_connection_source_direct() {
         let source = ConnectionSource::Direct {