Add support for VACUUM in Redshift (apache#2005)

Author: yoavcloud

src/ast/mod.rs

@@ -4438,6 +4438,13 @@ pub enum Statement {
     /// ```
     /// [Snowflake](https://docs.snowflake.com/en/sql-reference/sql/create-user)
     CreateUser(CreateUser),
+    /// Re-sorts rows and reclaims space in either a specified table or all tables in the current database
+    ///
+    /// ```sql
+    /// VACUUM tbl
+    /// ```
+    /// [Redshift](https://docs.aws.amazon.com/redshift/latest/dg/r_VACUUM_command.html)
+    Vacuum(VacuumStatement),
 }
 
 /// ```sql
@@ -6372,6 +6379,7 @@ impl fmt::Display for Statement {
             Statement::ExportData(e) => write!(f, "{e}"),
             Statement::CreateUser(s) => write!(f, "{s}"),
             Statement::AlterSchema(s) => write!(f, "{s}"),
+            Statement::Vacuum(s) => write!(f, "{s}"),
         }
     }
 }
@@ -10633,6 +10641,50 @@ impl fmt::Display for InitializeKind {
     }
 }
 
+/// Re-sorts rows and reclaims space in either a specified table or all tables in the current database
+///
+/// '''sql
+/// VACUUM [ FULL | SORT ONLY | DELETE ONLY | REINDEX | RECLUSTER ] [ \[ table_name \] [ TO threshold PERCENT ] \[ BOOST \] ]
+/// '''
+/// [Redshift](https://docs.aws.amazon.com/redshift/latest/dg/r_VACUUM_command.html)
+#[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
+#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
+#[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
+pub struct VacuumStatement {
+    pub full: bool,
+    pub sort_only: bool,
+    pub delete_only: bool,
+    pub reindex: bool,
+    pub recluster: bool,
+    pub table_name: Option<ObjectName>,
+    pub threshold: Option<Value>,
+    pub boost: bool,
+}
+
+impl fmt::Display for VacuumStatement {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        write!(
+            f,
+            "VACUUM{}{}{}{}{}",
+            if self.full { " FULL" } else { "" },
+            if self.sort_only { " SORT ONLY" } else { "" },
+            if self.delete_only { " DELETE ONLY" } else { "" },
+            if self.reindex { " REINDEX" } else { "" },
+            if self.recluster { " RECLUSTER" } else { "" },
+        )?;
+        if let Some(table_name) = &self.table_name {
+            write!(f, " {table_name}")?;
+        }
+        if let Some(threshold) = &self.threshold {
+            write!(f, " TO {threshold} PERCENT")?;
+        }
+        if self.boost {
+            write!(f, " BOOST")?;
+        }
+        Ok(())
+    }
+}
+
 #[cfg(test)]
 mod tests {
     use crate::tokenizer::Location;

src/ast/spans.rs

@@ -552,6 +552,7 @@ impl Spanned for Statement {
             ),
             Statement::CreateUser(..) => Span::empty(),
             Statement::AlterSchema(s) => s.span(),
+            Statement::Vacuum(..) => Span::empty(),
         }
     }
 }

src/keywords.rs

@@ -144,6 +144,7 @@ define_keywords!(
     BLOOMFILTER,
     BOOL,
     BOOLEAN,
+    BOOST,
     BOTH,
     BOX,
     BRIN,
@@ -761,6 +762,7 @@ define_keywords!(
     REGR_SXX,
     REGR_SXY,
     REGR_SYY,
+    REINDEX,
     RELATIVE,
     RELAY,
     RELEASE,

src/parser/mod.rs

@@ -649,6 +649,10 @@ impl<'a> Parser<'a> {
                     self.prev_token();
                     self.parse_export_data()
                 }
+                Keyword::VACUUM => {
+                    self.prev_token();
+                    self.parse_vacuum()
+                }
                 _ => self.expected("an SQL statement", next_token),
             },
             Token::LParen => {
@@ -16964,6 +16968,40 @@ impl<'a> Parser<'a> {
         }))
     }
 
+    fn parse_vacuum(&mut self) -> Result<Statement, ParserError> {
+        self.expect_keyword(Keyword::VACUUM)?;
+        let full = self.parse_keyword(Keyword::FULL);
+        let sort_only = self.parse_keywords(&[Keyword::SORT, Keyword::ONLY]);
+        let delete_only = self.parse_keywords(&[Keyword::DELETE, Keyword::ONLY]);
+        let reindex = self.parse_keyword(Keyword::REINDEX);
+        let recluster = self.parse_keyword(Keyword::RECLUSTER);
+        let (table_name, threshold, boost) =
+            match self.maybe_parse(|p| p.parse_object_name(false))? {
+                Some(table_name) => {
+                    let threshold = if self.parse_keyword(Keyword::TO) {
+                        let value = self.parse_value()?;
+                        self.expect_keyword(Keyword::PERCENT)?;
+                        Some(value.value)
+                    } else {
+                        None
+                    };
+                    let boost = self.parse_keyword(Keyword::BOOST);
+                    (Some(table_name), threshold, boost)
+                }
+                _ => (None, None, false),
+            };
+        Ok(Statement::Vacuum(VacuumStatement {
+            full,
+            sort_only,
+            delete_only,
+            reindex,
+            recluster,
+            table_name,
+            threshold,
+            boost,
+        }))
+    }
+
     /// Consume the parser and return its underlying token buffer
     pub fn into_tokens(self) -> Vec<TokenWithSpan> {
         self.tokens

tests/sqlparser_redshift.rs

@@ -413,3 +413,48 @@ fn parse_string_literal_backslash_escape() {
 fn parse_utf8_multibyte_idents() {
     redshift().verified_stmt("SELECT 🚀.city AS 🎸 FROM customers AS 🚀");
 }
+
+#[test]
+fn parse_vacuum() {
+    let stmt = redshift().verified_stmt("VACUUM FULL");
+    match stmt {
+        Statement::Vacuum(v) => {
+            assert!(v.full);
+            assert_eq!(v.table_name, None);
+        }
+        _ => unreachable!(),
+    }
+    let stmt = redshift().verified_stmt("VACUUM tbl");
+    match stmt {
+        Statement::Vacuum(v) => {
+            assert_eq!(
+                v.table_name,
+                Some(ObjectName::from(vec![Ident::new("tbl"),]))
+            );
+        }
+        _ => unreachable!(),
+    }
+    let stmt = redshift().verified_stmt(
+        "VACUUM FULL SORT ONLY DELETE ONLY REINDEX RECLUSTER db1.sc1.tbl1 TO 20 PERCENT BOOST",
+    );
+    match stmt {
+        Statement::Vacuum(v) => {
+            assert!(v.full);
+            assert!(v.sort_only);
+            assert!(v.delete_only);
+            assert!(v.reindex);
+            assert!(v.recluster);
+            assert_eq!(
+                v.table_name,
+                Some(ObjectName::from(vec![
+                    Ident::new("db1"),
+                    Ident::new("sc1"),
+                    Ident::new("tbl1"),
+                ]))
+            );
+            assert_eq!(v.threshold, Some(number("20")));
+            assert!(v.boost);
+        }
+        _ => unreachable!(),
+    }
+}