Support ADD PROJECTION syntax for ClickHouse (apache#1390)

Author: git-hulk

src/ast/ddl.rs

@@ -26,7 +26,7 @@ use sqlparser_derive::{Visit, VisitMut};
 use crate::ast::value::escape_single_quote_string;
 use crate::ast::{
     display_comma_separated, display_separated, DataType, Expr, Ident, MySQLColumnPosition,
-    ObjectName, SequenceOptions, SqlOption,
+    ObjectName, ProjectionSelect, SequenceOptions, SqlOption,
 };
 use crate::tokenizer::Token;
 
@@ -48,6 +48,15 @@ pub enum AlterTableOperation {
         /// MySQL `ALTER TABLE` only  [FIRST | AFTER column_name]
         column_position: Option<MySQLColumnPosition>,
     },
+    /// `ADD PROJECTION [IF NOT EXISTS] name ( SELECT <COLUMN LIST EXPR> [GROUP BY] [ORDER BY])`
+    ///
+    /// Note: this is a ClickHouse-specific operation.
+    /// Please refer to [ClickHouse](https://clickhouse.com/docs/en/sql-reference/statements/alter/projection#add-projection)
+    AddProjection {
+        if_not_exists: bool,
+        name: Ident,
+        select: ProjectionSelect,
+    },
     /// `DISABLE ROW LEVEL SECURITY`
     ///
     /// Note: this is a PostgreSQL-specific operation.
@@ -255,6 +264,17 @@ impl fmt::Display for AlterTableOperation {
 
                 Ok(())
             }
+            AlterTableOperation::AddProjection {
+                if_not_exists,
+                name,
+                select: query,
+            } => {
+                write!(f, "ADD PROJECTION")?;
+                if *if_not_exists {
+                    write!(f, " IF NOT EXISTS")?;
+                }
+                write!(f, " {} ({})", name, query)
+            }
             AlterTableOperation::AlterColumn { column_name, op } => {
                 write!(f, "ALTER COLUMN {column_name} {op}")
             }

src/ast/mod.rs

@@ -48,7 +48,7 @@ pub use self::query::{
     InterpolateExpr, Join, JoinConstraint, JoinOperator, JsonTableColumn,
     JsonTableColumnErrorHandling, LateralView, LockClause, LockType, MatchRecognizePattern,
     MatchRecognizeSymbol, Measure, NamedWindowDefinition, NamedWindowExpr, NonBlock, Offset,
-    OffsetRows, OrderBy, OrderByExpr, PivotValueSource, Query, RenameSelectItem,
+    OffsetRows, OrderBy, OrderByExpr, PivotValueSource, ProjectionSelect, Query, RenameSelectItem,
     RepetitionQuantifier, ReplaceSelectElement, ReplaceSelectItem, RowsPerMatch, Select,
     SelectInto, SelectItem, SetExpr, SetOperator, SetQuantifier, Setting, SymbolDefinition, Table,
     TableAlias, TableFactor, TableFunctionArgs, TableVersion, TableWithJoins, Top, TopQuantity,

src/ast/query.rs

@@ -68,16 +68,7 @@ impl fmt::Display for Query {
         }
         write!(f, "{}", self.body)?;
         if let Some(ref order_by) = self.order_by {
-            write!(f, " ORDER BY")?;
-            if !order_by.exprs.is_empty() {
-                write!(f, " {}", display_comma_separated(&order_by.exprs))?;
-            }
-            if let Some(ref interpolate) = order_by.interpolate {
-                match &interpolate.exprs {
-                    Some(exprs) => write!(f, " INTERPOLATE ({})", display_comma_separated(exprs))?,
-                    None => write!(f, " INTERPOLATE")?,
-                }
-            }
+            write!(f, " {order_by}")?;
         }
         if let Some(ref limit) = self.limit {
             write!(f, " LIMIT {limit}")?;
@@ -107,6 +98,33 @@ impl fmt::Display for Query {
     }
 }
 
+/// Query syntax for ClickHouse ADD PROJECTION statement.
+/// Its syntax is similar to SELECT statement, but it is used to add a new projection to a table.
+/// Syntax is `SELECT <COLUMN LIST EXPR> [GROUP BY] [ORDER BY]`
+///
+/// [ClickHouse](https://clickhouse.com/docs/en/sql-reference/statements/alter/projection#add-projection)
+#[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
+#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
+#[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
+pub struct ProjectionSelect {
+    pub projection: Vec<SelectItem>,
+    pub order_by: Option<OrderBy>,
+    pub group_by: Option<GroupByExpr>,
+}
+
+impl fmt::Display for ProjectionSelect {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        write!(f, "SELECT {}", display_comma_separated(&self.projection))?;
+        if let Some(ref group_by) = self.group_by {
+            write!(f, " {group_by}")?;
+        }
+        if let Some(ref order_by) = self.order_by {
+            write!(f, " {order_by}")?;
+        }
+        Ok(())
+    }
+}
+
 /// A node in a tree, representing a "query body" expression, roughly:
 /// `SELECT ... [ {UNION|EXCEPT|INTERSECT} SELECT ...]`
 #[allow(clippy::large_enum_variant)]
@@ -1729,6 +1747,22 @@ pub struct OrderBy {
     pub interpolate: Option<Interpolate>,
 }
 
+impl fmt::Display for OrderBy {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        write!(f, "ORDER BY")?;
+        if !self.exprs.is_empty() {
+            write!(f, " {}", display_comma_separated(&self.exprs))?;
+        }
+        if let Some(ref interpolate) = self.interpolate {
+            match &interpolate.exprs {
+                Some(exprs) => write!(f, " INTERPOLATE ({})", display_comma_separated(exprs))?,
+                None => write!(f, " INTERPOLATE")?,
+            }
+        }
+        Ok(())
+    }
+}
+
 /// An `ORDER BY` expression
 #[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]

src/keywords.rs

@@ -576,6 +576,7 @@ define_keywords!(
     PRIVILEGES,
     PROCEDURE,
     PROGRAM,
+    PROJECTION,
     PURGE,
     QUALIFY,
     QUARTER,

src/parser/mod.rs

@@ -6438,10 +6438,38 @@ impl<'a> Parser<'a> {
         Ok(Partition::Partitions(partitions))
     }
 
+    pub fn parse_projection_select(&mut self) -> Result<ProjectionSelect, ParserError> {
+        self.expect_token(&Token::LParen)?;
+        self.expect_keyword(Keyword::SELECT)?;
+        let projection = self.parse_projection()?;
+        let group_by = self.parse_optional_group_by()?;
+        let order_by = self.parse_optional_order_by()?;
+        self.expect_token(&Token::RParen)?;
+        Ok(ProjectionSelect {
+            projection,
+            group_by,
+            order_by,
+        })
+    }
+    pub fn parse_alter_table_add_projection(&mut self) -> Result<AlterTableOperation, ParserError> {
+        let if_not_exists = self.parse_keywords(&[Keyword::IF, Keyword::NOT, Keyword::EXISTS]);
+        let name = self.parse_identifier(false)?;
+        let query = self.parse_projection_select()?;
+        Ok(AlterTableOperation::AddProjection {
+            if_not_exists,
+            name,
+            select: query,
+        })
+    }
+
     pub fn parse_alter_table_operation(&mut self) -> Result<AlterTableOperation, ParserError> {
         let operation = if self.parse_keyword(Keyword::ADD) {
             if let Some(constraint) = self.parse_optional_table_constraint()? {
                 AlterTableOperation::AddConstraint(constraint)
+            } else if dialect_of!(self is ClickHouseDialect|GenericDialect)
+                && self.parse_keyword(Keyword::PROJECTION)
+            {
+                return self.parse_alter_table_add_projection();
             } else {
                 let if_not_exists =
                     self.parse_keywords(&[Keyword::IF, Keyword::NOT, Keyword::EXISTS]);
@@ -7686,6 +7714,66 @@ impl<'a> Parser<'a> {
         }
     }
 
+    pub fn parse_optional_group_by(&mut self) -> Result<Option<GroupByExpr>, ParserError> {
+        if self.parse_keywords(&[Keyword::GROUP, Keyword::BY]) {
+            let expressions = if self.parse_keyword(Keyword::ALL) {
+                None
+            } else {
+                Some(self.parse_comma_separated(Parser::parse_group_by_expr)?)
+            };
+
+            let mut modifiers = vec![];
+            if dialect_of!(self is ClickHouseDialect | GenericDialect) {
+                loop {
+                    if !self.parse_keyword(Keyword::WITH) {
+                        break;
+                    }
+                    let keyword = self.expect_one_of_keywords(&[
+                        Keyword::ROLLUP,
+                        Keyword::CUBE,
+                        Keyword::TOTALS,
+                    ])?;
+                    modifiers.push(match keyword {
+                        Keyword::ROLLUP => GroupByWithModifier::Rollup,
+                        Keyword::CUBE => GroupByWithModifier::Cube,
+                        Keyword::TOTALS => GroupByWithModifier::Totals,
+                        _ => {
+                            return parser_err!(
+                                "BUG: expected to match GroupBy modifier keyword",
+                                self.peek_token().location
+                            )
+                        }
+                    });
+                }
+            }
+            let group_by = match expressions {
+                None => GroupByExpr::All(modifiers),
+                Some(exprs) => GroupByExpr::Expressions(exprs, modifiers),
+            };
+            Ok(Some(group_by))
+        } else {
+            Ok(None)
+        }
+    }
+
+    pub fn parse_optional_order_by(&mut self) -> Result<Option<OrderBy>, ParserError> {
+        if self.parse_keywords(&[Keyword::ORDER, Keyword::BY]) {
+            let order_by_exprs = self.parse_comma_separated(Parser::parse_order_by_expr)?;
+            let interpolate = if dialect_of!(self is ClickHouseDialect | GenericDialect) {
+                self.parse_interpolations()?
+            } else {
+                None
+            };
+
+            Ok(Some(OrderBy {
+                exprs: order_by_exprs,
+                interpolate,
+            }))
+        } else {
+            Ok(None)
+        }
+    }
+
     /// Parse a possibly qualified, possibly quoted identifier, e.g.
     /// `foo` or `myschema."table"
     ///
@@ -8278,21 +8366,7 @@ impl<'a> Parser<'a> {
         } else {
             let body = self.parse_boxed_query_body(self.dialect.prec_unknown())?;
 
-            let order_by = if self.parse_keywords(&[Keyword::ORDER, Keyword::BY]) {
-                let order_by_exprs = self.parse_comma_separated(Parser::parse_order_by_expr)?;
-                let interpolate = if dialect_of!(self is ClickHouseDialect | GenericDialect) {
-                    self.parse_interpolations()?
-                } else {
-                    None
-                };
-
-                Some(OrderBy {
-                    exprs: order_by_exprs,
-                    interpolate,
-                })
-            } else {
-                None
-            };
+            let order_by = self.parse_optional_order_by()?;
 
             let mut limit = None;
             let mut offset = None;
@@ -8760,44 +8834,9 @@ impl<'a> Parser<'a> {
             None
         };
 
-        let group_by = if self.parse_keywords(&[Keyword::GROUP, Keyword::BY]) {
-            let expressions = if self.parse_keyword(Keyword::ALL) {
-                None
-            } else {
-                Some(self.parse_comma_separated(Parser::parse_group_by_expr)?)
-            };
-
-            let mut modifiers = vec![];
-            if dialect_of!(self is ClickHouseDialect | GenericDialect) {
-                loop {
-                    if !self.parse_keyword(Keyword::WITH) {
-                        break;
-                    }
-                    let keyword = self.expect_one_of_keywords(&[
-                        Keyword::ROLLUP,
-                        Keyword::CUBE,
-                        Keyword::TOTALS,
-                    ])?;
-                    modifiers.push(match keyword {
-                        Keyword::ROLLUP => GroupByWithModifier::Rollup,
-                        Keyword::CUBE => GroupByWithModifier::Cube,
-                        Keyword::TOTALS => GroupByWithModifier::Totals,
-                        _ => {
-                            return parser_err!(
-                                "BUG: expected to match GroupBy modifier keyword",
-                                self.peek_token().location
-                            )
-                        }
-                    });
-                }
-            }
-            match expressions {
-                None => GroupByExpr::All(modifiers),
-                Some(exprs) => GroupByExpr::Expressions(exprs, modifiers),
-            }
-        } else {
-            GroupByExpr::Expressions(vec![], vec![])
-        };
+        let group_by = self
+            .parse_optional_group_by()?
+            .unwrap_or_else(|| GroupByExpr::Expressions(vec![], vec![]));
 
         let cluster_by = if self.parse_keywords(&[Keyword::CLUSTER, Keyword::BY]) {
             self.parse_comma_separated(Parser::parse_expr)?

tests/sqlparser_clickhouse.rs

@@ -287,6 +287,78 @@ fn parse_alter_table_attach_and_detach_partition() {
     }
 }
 
+#[test]
+fn parse_alter_table_add_projection() {
+    match clickhouse_and_generic().verified_stmt(concat!(
+        "ALTER TABLE t0 ADD PROJECTION IF NOT EXISTS my_name",
+        " (SELECT a, b GROUP BY a ORDER BY b)",
+    )) {
+        Statement::AlterTable {
+            name, operations, ..
+        } => {
+            assert_eq!(name, ObjectName(vec!["t0".into()]));
+            assert_eq!(1, operations.len());
+            assert_eq!(
+                operations[0],
+                AlterTableOperation::AddProjection {
+                    if_not_exists: true,
+                    name: "my_name".into(),
+                    select: ProjectionSelect {
+                        projection: vec![
+                            UnnamedExpr(Identifier(Ident::new("a"))),
+                            UnnamedExpr(Identifier(Ident::new("b"))),
+                        ],
+                        group_by: Some(GroupByExpr::Expressions(
+                            vec![Identifier(Ident::new("a"))],
+                            vec![]
+                        )),
+                        order_by: Some(OrderBy {
+                            exprs: vec![OrderByExpr {
+                                expr: Identifier(Ident::new("b")),
+                                asc: None,
+                                nulls_first: None,
+                                with_fill: None,
+                            }],
+                            interpolate: None,
+                        }),
+                    }
+                }
+            )
+        }
+        _ => unreachable!(),
+    }
+
+    // leave out IF NOT EXISTS is allowed
+    clickhouse_and_generic()
+        .verified_stmt("ALTER TABLE t0 ADD PROJECTION my_name (SELECT a, b GROUP BY a ORDER BY b)");
+    // leave out GROUP BY is allowed
+    clickhouse_and_generic()
+        .verified_stmt("ALTER TABLE t0 ADD PROJECTION my_name (SELECT a, b ORDER BY b)");
+    // leave out ORDER BY is allowed
+    clickhouse_and_generic()
+        .verified_stmt("ALTER TABLE t0 ADD PROJECTION my_name (SELECT a, b GROUP BY a)");
+
+    // missing select query is not allowed
+    assert_eq!(
+        clickhouse_and_generic()
+            .parse_sql_statements("ALTER TABLE t0 ADD PROJECTION my_name")
+            .unwrap_err(),
+        ParserError("Expected: (, found: EOF".to_string())
+    );
+    assert_eq!(
+        clickhouse_and_generic()
+            .parse_sql_statements("ALTER TABLE t0 ADD PROJECTION my_name ()")
+            .unwrap_err(),
+        ParserError("Expected: SELECT, found: )".to_string())
+    );
+    assert_eq!(
+        clickhouse_and_generic()
+            .parse_sql_statements("ALTER TABLE t0 ADD PROJECTION my_name (SELECT)")
+            .unwrap_err(),
+        ParserError("Expected: an expression:, found: )".to_string())
+    );
+}
+
 #[test]
 fn parse_optimize_table() {
     clickhouse_and_generic().verified_stmt("OPTIMIZE TABLE t0");