Support COPY TO Externally Defined File Formats, add FileType trait

datafusion-cli/src/exec.rs

@@ -21,7 +21,6 @@ use std::collections::HashMap;
 use std::fs::File;
 use std::io::prelude::*;
 use std::io::BufReader;
-use std::str::FromStr;
 
 use crate::cli_context::CliSessionContext;
 use crate::helper::split_from_semicolon;
@@ -35,14 +34,14 @@ use crate::{
 
 use datafusion::common::instant::Instant;
 use datafusion::common::plan_datafusion_err;
+use datafusion::config::ConfigFileType;
 use datafusion::datasource::listing::ListingTableUrl;
 use datafusion::error::{DataFusionError, Result};
 use datafusion::logical_expr::{DdlStatement, LogicalPlan};
 use datafusion::physical_plan::{collect, execute_stream, ExecutionPlanProperties};
 use datafusion::sql::parser::{DFParser, Statement};
 use datafusion::sql::sqlparser::dialect::dialect_from_str;
 
-use datafusion::common::FileType;
 use datafusion::sql::sqlparser;
 use rustyline::error::ReadlineError;
 use rustyline::Editor;
@@ -291,6 +290,15 @@ impl AdjustedPrintOptions {
     }
 }
 
+fn config_file_type_from_str(ext: &str) -> Option<ConfigFileType> {
+    match ext.to_lowercase().as_str() {
+        "csv" => Some(ConfigFileType::CSV),
+        "json" => Some(ConfigFileType::JSON),
+        "parquet" => Some(ConfigFileType::PARQUET),
+        _ => None,
+    }
+}
+
 async fn create_plan(
     ctx: &mut dyn CliSessionContext,
     statement: Statement,
@@ -302,7 +310,7 @@ async fn create_plan(
     // will raise Configuration errors.
     if let LogicalPlan::Ddl(DdlStatement::CreateExternalTable(cmd)) = &plan {
         // To support custom formats, treat error as None
-        let format = FileType::from_str(&cmd.file_type).ok();
+        let format = config_file_type_from_str(&cmd.file_type);
         register_object_store_and_config_extensions(
             ctx,
             &cmd.location,
@@ -313,13 +321,13 @@ async fn create_plan(
     }
 
     if let LogicalPlan::Copy(copy_to) = &mut plan {
-        let format: FileType = (&copy_to.format_options).into();
+        let format = config_file_type_from_str(&copy_to.file_type.get_ext());
 
         register_object_store_and_config_extensions(
             ctx,
             &copy_to.output_url,
             &copy_to.options,
-            Some(format),
+            format,
         )
         .await?;
     }
@@ -357,7 +365,7 @@ pub(crate) async fn register_object_store_and_config_extensions(
     ctx: &dyn CliSessionContext,
     location: &String,
     options: &HashMap<String, String>,
-    format: Option<FileType>,
+    format: Option<ConfigFileType>,
 ) -> Result<()> {
     // Parse the location URL to extract the scheme and other components
     let table_path = ListingTableUrl::parse(location)?;
@@ -392,7 +400,6 @@ pub(crate) async fn register_object_store_and_config_extensions(
 mod tests {
     use super::*;
 
-    use datafusion::common::config::FormatOptions;
     use datafusion::common::plan_err;
 
     use datafusion::prelude::SessionContext;
@@ -403,7 +410,7 @@ mod tests {
         let plan = ctx.state().create_logical_plan(sql).await?;
 
         if let LogicalPlan::Ddl(DdlStatement::CreateExternalTable(cmd)) = &plan {
-            let format = FileType::from_str(&cmd.file_type).ok();
+            let format = config_file_type_from_str(&cmd.file_type);
             register_object_store_and_config_extensions(
                 &ctx,
                 &cmd.location,
@@ -429,12 +436,12 @@ mod tests {
         let plan = ctx.state().create_logical_plan(sql).await?;
 
         if let LogicalPlan::Copy(cmd) = &plan {
-            let format: FileType = (&cmd.format_options).into();
+            let format = config_file_type_from_str(&cmd.file_type.get_ext());
             register_object_store_and_config_extensions(
                 &ctx,
                 &cmd.output_url,
                 &cmd.options,
-                Some(format),
+                format,
             )
             .await?;
         } else {
@@ -484,7 +491,7 @@ mod tests {
                 let mut plan = create_plan(&mut ctx, statement).await?;
                 if let LogicalPlan::Copy(copy_to) = &mut plan {
                     assert_eq!(copy_to.output_url, location);
-                    assert!(matches!(copy_to.format_options, FormatOptions::PARQUET(_)));
+                    assert_eq!(copy_to.file_type.get_ext(), "parquet".to_string());
                     ctx.runtime_env()
                         .object_store_registry
                         .get_store(&Url::parse(&copy_to.output_url).unwrap())?;

datafusion-examples/examples/external_dependency/dataframe-to-s3.rs

@@ -20,10 +20,10 @@ use std::sync::Arc;
 
 use datafusion::dataframe::DataFrameWriteOptions;
 use datafusion::datasource::file_format::parquet::ParquetFormat;
+use datafusion::datasource::file_format::FileFormat;
 use datafusion::datasource::listing::ListingOptions;
 use datafusion::error::Result;
 use datafusion::prelude::*;
-use datafusion_common::{FileType, GetExt};
 
 use object_store::aws::AmazonS3Builder;
 use url::Url;
@@ -54,7 +54,7 @@ async fn main() -> Result<()> {
     let path = format!("s3://{bucket_name}/test_data/");
     let file_format = ParquetFormat::default().with_enable_pruning(true);
     let listing_options = ListingOptions::new(Arc::new(file_format))
-        .with_file_extension(FileType::PARQUET.get_ext());
+        .with_file_extension(ParquetFormat::default().get_ext());
     ctx.register_listing_table("test", &path, listing_options, None, None)
         .await?;
 
@@ -79,7 +79,7 @@ async fn main() -> Result<()> {
 
     let file_format = ParquetFormat::default().with_enable_pruning(true);
     let listing_options = ListingOptions::new(Arc::new(file_format))
-        .with_file_extension(FileType::PARQUET.get_ext());
+        .with_file_extension(ParquetFormat::default().get_ext());
     ctx.register_listing_table("test2", &out_path, listing_options, None, None)
         .await?;
 

datafusion/common/src/config.rs

@@ -24,7 +24,7 @@ use std::str::FromStr;
 
 use crate::error::_config_err;
 use crate::parsers::CompressionTypeVariant;
-use crate::{DataFusionError, FileType, Result};
+use crate::{DataFusionError, Result};
 
 /// A macro that wraps a configuration struct and automatically derives
 /// [`Default`] and [`ConfigField`] for it, allowing it to be used
@@ -1116,6 +1116,16 @@ macro_rules! extensions_options {
     }
 }
 
+/// These file types have special built in behavior for configuration.
+/// Use TableOptions::Extensions for configuring other file types.
+#[derive(Debug, Clone)]
+pub enum ConfigFileType {
+    CSV,
+    #[cfg(feature = "parquet")]
+    PARQUET,
+    JSON,
+}
+
 /// Represents the configuration options available for handling different table formats within a data processing application.
 /// This struct encompasses options for various file formats including CSV, Parquet, and JSON, allowing for flexible configuration
 /// of parsing and writing behaviors specific to each format. Additionally, it supports extending functionality through custom extensions.
@@ -1134,7 +1144,7 @@ pub struct TableOptions {
 
     /// The current file format that the table operations should assume. This option allows
     /// for dynamic switching between the supported file types (e.g., CSV, Parquet, JSON).
-    pub current_format: Option<FileType>,
+    pub current_format: Option<ConfigFileType>,
 
     /// Optional extensions that can be used to extend or customize the behavior of the table
     /// options. Extensions can be registered using `Extensions::insert` and might include
@@ -1152,10 +1162,9 @@ impl ConfigField for TableOptions {
         if let Some(file_type) = &self.current_format {
             match file_type {
                 #[cfg(feature = "parquet")]
-                FileType::PARQUET => self.parquet.visit(v, "format", ""),
-                FileType::CSV => self.csv.visit(v, "format", ""),
-                FileType::JSON => self.json.visit(v, "format", ""),
-                _ => {}
+                ConfigFileType::PARQUET => self.parquet.visit(v, "format", ""),
+                ConfigFileType::CSV => self.csv.visit(v, "format", ""),
+                ConfigFileType::JSON => self.json.visit(v, "format", ""),
             }
         } else {
             self.csv.visit(v, "csv", "");
@@ -1188,12 +1197,9 @@ impl ConfigField for TableOptions {
         match key {
             "format" => match format {
                 #[cfg(feature = "parquet")]
-                FileType::PARQUET => self.parquet.set(rem, value),
-                FileType::CSV => self.csv.set(rem, value),
-                FileType::JSON => self.json.set(rem, value),
-                _ => {
-                    _config_err!("Config value \"{key}\" is not supported on {}", format)
-                }
+                ConfigFileType::PARQUET => self.parquet.set(rem, value),
+                ConfigFileType::CSV => self.csv.set(rem, value),
+                ConfigFileType::JSON => self.json.set(rem, value),
             },
             _ => _config_err!("Config value \"{key}\" not found on TableOptions"),
         }
@@ -1210,15 +1216,6 @@ impl TableOptions {
         Self::default()
     }
 
-    /// Sets the file format for the table.
-    ///
-    /// # Parameters
-    ///
-    /// * `format`: The file format to use (e.g., CSV, Parquet).
-    pub fn set_file_format(&mut self, format: FileType) {
-        self.current_format = Some(format);
-    }
-
     /// Creates a new `TableOptions` instance initialized with settings from a given session config.
     ///
     /// # Parameters
@@ -1249,6 +1246,15 @@ impl TableOptions {
         clone
     }
 
+    /// Sets the file format for the table.
+    ///
+    /// # Parameters
+    ///
+    /// * `format`: The file format to use (e.g., CSV, Parquet).
+    pub fn set_file_format(&mut self, format: ConfigFileType) {
+        self.current_format = Some(format);
+    }
+
     /// Sets the extensions for this `TableOptions` instance.
     ///
     /// # Parameters
@@ -1658,6 +1664,8 @@ config_namespace! {
     }
 }
 
+pub trait FormatOptionsExt: Display {}
+
 #[derive(Debug, Clone, PartialEq)]
 #[allow(clippy::large_enum_variant)]
 pub enum FormatOptions {
@@ -1682,28 +1690,15 @@ impl Display for FormatOptions {
     }
 }
 
-impl From<FileType> for FormatOptions {
-    fn from(value: FileType) -> Self {
-        match value {
-            FileType::ARROW => FormatOptions::ARROW,
-            FileType::AVRO => FormatOptions::AVRO,
-            #[cfg(feature = "parquet")]
-            FileType::PARQUET => FormatOptions::PARQUET(TableParquetOptions::default()),
-            FileType::CSV => FormatOptions::CSV(CsvOptions::default()),
-            FileType::JSON => FormatOptions::JSON(JsonOptions::default()),
-        }
-    }
-}
-
 #[cfg(test)]
 mod tests {
     use std::any::Any;
     use std::collections::HashMap;
 
     use crate::config::{
-        ConfigEntry, ConfigExtension, ExtensionOptions, Extensions, TableOptions,
+        ConfigEntry, ConfigExtension, ConfigFileType, ExtensionOptions, Extensions,
+        TableOptions,
     };
-    use crate::FileType;
 
     #[derive(Default, Debug, Clone)]
     pub struct TestExtensionConfig {
@@ -1761,7 +1756,7 @@ mod tests {
         let mut extension = Extensions::new();
         extension.insert(TestExtensionConfig::default());
         let mut table_config = TableOptions::new().with_extensions(extension);
-        table_config.set_file_format(FileType::CSV);
+        table_config.set_file_format(ConfigFileType::CSV);
         table_config.set("format.delimiter", ";").unwrap();
         assert_eq!(table_config.csv.delimiter, b';');
         table_config.set("test.bootstrap.servers", "asd").unwrap();
@@ -1778,7 +1773,7 @@ mod tests {
     #[test]
     fn csv_u8_table_options() {
         let mut table_config = TableOptions::new();
-        table_config.set_file_format(FileType::CSV);
+        table_config.set_file_format(ConfigFileType::CSV);
         table_config.set("format.delimiter", ";").unwrap();
         assert_eq!(table_config.csv.delimiter as char, ';');
         table_config.set("format.escape", "\"").unwrap();
@@ -1791,7 +1786,7 @@ mod tests {
     #[test]
     fn parquet_table_options() {
         let mut table_config = TableOptions::new();
-        table_config.set_file_format(FileType::PARQUET);
+        table_config.set_file_format(ConfigFileType::PARQUET);
         table_config
             .set("format.bloom_filter_enabled::col1", "true")
             .unwrap();
@@ -1805,7 +1800,7 @@ mod tests {
     #[test]
     fn parquet_table_options_config_entry() {
         let mut table_config = TableOptions::new();
-        table_config.set_file_format(FileType::PARQUET);
+        table_config.set_file_format(ConfigFileType::PARQUET);
         table_config
             .set("format.bloom_filter_enabled::col1", "true")
             .unwrap();
@@ -1819,7 +1814,7 @@ mod tests {
     #[test]
     fn parquet_table_options_config_metadata_entry() {
         let mut table_config = TableOptions::new();
-        table_config.set_file_format(FileType::PARQUET);
+        table_config.set_file_format(ConfigFileType::PARQUET);
         table_config.set("format.metadata::key1", "").unwrap();
         table_config.set("format.metadata::key2", "value2").unwrap();
         table_config