sigma: support for parsing @sigma table factors and expressions (#19)

Author: jmhain

src/ast/mod.rs

@@ -374,6 +374,8 @@ pub enum Expr {
     Identifier(Ident),
     /// Multi-part identifier, e.g. `table_alias.column` or `schema.table.col`
     CompoundIdentifier(Vec<Ident>),
+    /// A reference to a Sigma scalar value, e.g. `@sigma.my_parameter`.
+    SigmaParameter(Ident),
     /// JSON access (postgres)  eg: data->'tags'
     JsonAccess {
         left: Box<Expr>,
@@ -752,6 +754,7 @@ impl fmt::Display for Expr {
     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
         match self {
             Expr::Identifier(s) => write!(f, "{s}"),
+            Expr::SigmaParameter(s) => write!(f, "@sigma.{s}"),
             Expr::MapAccess { column, keys } => {
                 write!(f, "{column}")?;
                 for k in keys {

src/ast/query.rs

@@ -761,6 +761,11 @@ pub enum TableFactor {
         /// [Partition selection](https://dev.mysql.com/doc/refman/8.0/en/partitioning-selection.html), supported by MySQL.
         partitions: Vec<Ident>,
     },
+    /// A reference to an element in a Sigma workbook.
+    SigmaElement {
+        element: Ident,
+        alias: Option<TableAlias>,
+    },
     Derived {
         lateral: bool,
         subquery: Box<Query>,
@@ -887,6 +892,13 @@ impl fmt::Display for TableFactor {
                 }
                 Ok(())
             }
+            TableFactor::SigmaElement { element, alias } => {
+                write!(f, "@sigma.{element}")?;
+                if let Some(alias) = alias {
+                    write!(f, " AS {alias}")?;
+                }
+                Ok(())
+            }
             TableFactor::Derived {
                 lateral,
                 subquery,

src/keywords.rs

@@ -612,6 +612,7 @@ define_keywords!(
     SETS,
     SHARE,
     SHOW,
+    SIGMA,
     SIMILAR,
     SKIP,
     SLOW,

src/parser/mod.rs

@@ -1050,6 +1050,10 @@ impl<'a> Parser<'a> {
             }, // End of Token::Word
             // array `[1, 2, 3]`
             Token::LBracket => self.parse_array_expr(false),
+            Token::AtSign if self.parse_keyword(Keyword::SIGMA) => {
+                self.expect_token(&Token::Period)?;
+                Ok(Expr::SigmaParameter(self.parse_identifier(false)?))
+            }
             tok @ Token::Minus | tok @ Token::Plus => {
                 let op = if tok == Token::Plus {
                     UnaryOperator::Plus
@@ -7985,6 +7989,7 @@ impl<'a> Parser<'a> {
                     match &mut table_and_joins.relation {
                         TableFactor::Derived { alias, .. }
                         | TableFactor::Table { alias, .. }
+                        | TableFactor::SigmaElement { alias, .. }
                         | TableFactor::Function { alias, .. }
                         | TableFactor::UNNEST { alias, .. }
                         | TableFactor::JsonTable { alias, .. }
@@ -8062,6 +8067,11 @@ impl<'a> Parser<'a> {
                 columns,
                 alias,
             })
+        } else if self.parse_sigma_directive() {
+            self.expect_token(&Token::Period)?;
+            let element = self.parse_identifier(true)?;
+            let alias = self.parse_optional_table_alias(keywords::RESERVED_FOR_TABLE_ALIAS)?;
+            Ok(TableFactor::SigmaElement { element, alias })
         } else {
             let name = self.parse_object_name(true)?;
 
@@ -8118,6 +8128,15 @@ impl<'a> Parser<'a> {
         }
     }
 
+    pub fn parse_sigma_directive(&mut self) -> bool {
+        self.maybe_parse(|p| {
+            p.expect_token(&Token::AtSign)?;
+            p.expect_keyword(Keyword::SIGMA)?;
+            Ok(())
+        })
+        .is_some()
+    }
+
     /// Parse a given table version specifier.
     ///
     /// For now it only supports timestamp versioning for BigQuery and MSSQL dialects.

tests/sqlparser_sigma.rs

@@ -0,0 +1,38 @@
+#![warn(clippy::all)]
+
+use sqlparser::ast::*;
+use sqlparser::dialect::SnowflakeDialect;
+use test_utils::*;
+
+#[macro_use]
+mod test_utils;
+
+fn snowflake() -> TestedDialects {
+    TestedDialects {
+        dialects: vec![Box::new(SnowflakeDialect {})],
+        options: None,
+    }
+}
+#[test]
+fn parse_sigma() {
+    let sql = "SELECT my_column FROM @sigma.my_element WHERE my_column <> @sigma.param_filter";
+    let select = snowflake().verified_only_select(sql);
+    assert_eq!(
+        select.from,
+        vec![TableWithJoins {
+            relation: TableFactor::SigmaElement {
+                element: Ident::new("my_element"),
+                alias: None
+            },
+            joins: vec![]
+        }]
+    );
+    assert_eq!(
+        select.selection,
+        Some(Expr::BinaryOp {
+            left: Box::new(Expr::Identifier(Ident::new("my_column"))),
+            op: BinaryOperator::NotEq,
+            right: Box::new(Expr::SigmaParameter(Ident::new("param_filter"))),
+        })
+    )
+}