Add support for Hive's LOAD DATA expr (apache#1520)

Co-authored-by: Ifeanyi Ubah <[email protected]>

Author: wugeer

src/ast/mod.rs

@@ -3347,6 +3347,22 @@ pub enum Statement {
         channel: Ident,
         payload: Option<String>,
     },
+    /// ```sql
+    /// LOAD DATA [LOCAL] INPATH 'filepath' [OVERWRITE] INTO TABLE tablename
+    /// [PARTITION (partcol1=val1, partcol2=val2 ...)]
+    /// [INPUTFORMAT 'inputformat' SERDE 'serde']
+    /// ```
+    /// Loading files into tables
+    ///
+    /// See Hive <https://cwiki.apache.org/confluence/pages/viewpage.action?pageId=27362036#LanguageManualDML-Loadingfilesintotables>
+    LoadData {
+        local: bool,
+        inpath: String,
+        overwrite: bool,
+        table_name: ObjectName,
+        partitioned: Option<Vec<Expr>>,
+        table_format: Option<HiveLoadDataFormat>,
+    },
 }
 
 impl fmt::Display for Statement {
@@ -3949,6 +3965,36 @@ impl fmt::Display for Statement {
                 Ok(())
             }
             Statement::CreateTable(create_table) => create_table.fmt(f),
+            Statement::LoadData {
+                local,
+                inpath,
+                overwrite,
+                table_name,
+                partitioned,
+                table_format,
+            } => {
+                write!(
+                    f,
+                    "LOAD DATA {local}INPATH '{inpath}' {overwrite}INTO TABLE {table_name}",
+                    local = if *local { "LOCAL " } else { "" },
+                    inpath = inpath,
+                    overwrite = if *overwrite { "OVERWRITE " } else { "" },
+                    table_name = table_name,
+                )?;
+                if let Some(ref parts) = &partitioned {
+                    if !parts.is_empty() {
+                        write!(f, " PARTITION ({})", display_comma_separated(parts))?;
+                    }
+                }
+                if let Some(HiveLoadDataFormat {
+                    serde,
+                    input_format,
+                }) = &table_format
+                {
+                    write!(f, " INPUTFORMAT {input_format} SERDE {serde}")?;
+                }
+                Ok(())
+            }
             Statement::CreateVirtualTable {
                 name,
                 if_not_exists,
@@ -5855,6 +5901,14 @@ pub enum HiveRowFormat {
     DELIMITED { delimiters: Vec<HiveRowDelimiter> },
 }
 
+#[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
+#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
+#[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
+pub struct HiveLoadDataFormat {
+    pub serde: Expr,
+    pub input_format: Expr,
+}
+
 #[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]

src/dialect/duckdb.rs

@@ -66,4 +66,9 @@ impl Dialect for DuckDbDialect {
     fn supports_explain_with_utility_options(&self) -> bool {
         true
     }
+
+    /// See DuckDB <https://duckdb.org/docs/sql/statements/load_and_install.html#load>
+    fn supports_load_extension(&self) -> bool {
+        true
+    }
 }

src/dialect/generic.rs

@@ -115,4 +115,8 @@ impl Dialect for GenericDialect {
     fn supports_comment_on(&self) -> bool {
         true
     }
+
+    fn supports_load_extension(&self) -> bool {
+        true
+    }
 }

src/dialect/hive.rs

@@ -56,4 +56,9 @@ impl Dialect for HiveDialect {
     fn supports_bang_not_operator(&self) -> bool {
         true
     }
+
+    /// See Hive <https://cwiki.apache.org/confluence/pages/viewpage.action?pageId=27362036#LanguageManualDML-Loadingfilesintotables>
+    fn supports_load_data(&self) -> bool {
+        true
+    }
 }

src/dialect/mod.rs

@@ -620,6 +620,16 @@ pub trait Dialect: Debug + Any {
         false
     }
 
+    /// Returns true if the dialect supports the `LOAD DATA` statement
+    fn supports_load_data(&self) -> bool {
+        false
+    }
+
+    /// Returns true if the dialect supports the `LOAD extension` statement
+    fn supports_load_extension(&self) -> bool {
+        false
+    }
+
     /// Returns true if this dialect expects the `TOP` option
     /// before the `ALL`/`DISTINCT` options in a `SELECT` statement.
     fn supports_top_before_distinct(&self) -> bool {

src/keywords.rs

@@ -389,6 +389,7 @@ define_keywords!(
     INITIALLY,
     INNER,
     INOUT,
+    INPATH,
     INPUT,
     INPUTFORMAT,
     INSENSITIVE,

src/parser/mod.rs

@@ -543,10 +543,7 @@ impl<'a> Parser<'a> {
                 Keyword::INSTALL if dialect_of!(self is DuckDbDialect | GenericDialect) => {
                     self.parse_install()
                 }
-                // `LOAD` is duckdb specific https://duckdb.org/docs/extensions/overview
-                Keyword::LOAD if dialect_of!(self is DuckDbDialect | GenericDialect) => {
-                    self.parse_load()
-                }
+                Keyword::LOAD => self.parse_load(),
                 // `OPTIMIZE` is clickhouse specific https://clickhouse.tech/docs/en/sql-reference/statements/optimize/
                 Keyword::OPTIMIZE if dialect_of!(self is ClickHouseDialect | GenericDialect) => {
                     self.parse_optimize_table()
@@ -11222,6 +11219,22 @@ impl<'a> Parser<'a> {
         }
     }
 
+    pub fn parse_load_data_table_format(
+        &mut self,
+    ) -> Result<Option<HiveLoadDataFormat>, ParserError> {
+        if self.parse_keyword(Keyword::INPUTFORMAT) {
+            let input_format = self.parse_expr()?;
+            self.expect_keyword(Keyword::SERDE)?;
+            let serde = self.parse_expr()?;
+            Ok(Some(HiveLoadDataFormat {
+                input_format,
+                serde,
+            }))
+        } else {
+            Ok(None)
+        }
+    }
+
     /// Parse an UPDATE statement, returning a `Box`ed SetExpr
     ///
     /// This is used to reduce the size of the stack frames in debug builds
@@ -12224,10 +12237,35 @@ impl<'a> Parser<'a> {
         Ok(Statement::Install { extension_name })
     }
 
-    /// `LOAD [extension_name]`
+    /// Parse a SQL LOAD statement
     pub fn parse_load(&mut self) -> Result<Statement, ParserError> {
-        let extension_name = self.parse_identifier(false)?;
-        Ok(Statement::Load { extension_name })
+        if self.dialect.supports_load_extension() {
+            let extension_name = self.parse_identifier(false)?;
+            Ok(Statement::Load { extension_name })
+        } else if self.parse_keyword(Keyword::DATA) && self.dialect.supports_load_data() {
+            let local = self.parse_one_of_keywords(&[Keyword::LOCAL]).is_some();
+            self.expect_keyword(Keyword::INPATH)?;
+            let inpath = self.parse_literal_string()?;
+            let overwrite = self.parse_one_of_keywords(&[Keyword::OVERWRITE]).is_some();
+            self.expect_keyword(Keyword::INTO)?;
+            self.expect_keyword(Keyword::TABLE)?;
+            let table_name = self.parse_object_name(false)?;
+            let partitioned = self.parse_insert_partition()?;
+            let table_format = self.parse_load_data_table_format()?;
+            Ok(Statement::LoadData {
+                local,
+                inpath,
+                overwrite,
+                table_name,
+                partitioned,
+                table_format,
+            })
+        } else {
+            self.expected(
+                "`DATA` or an extension name after `LOAD`",
+                self.peek_token(),
+            )
+        }
     }
 
     /// ```sql