Support: Databricks and generic: fix for values as table name (apache#1278)

Author: jmhain

src/ast/query.rs

@@ -108,6 +108,17 @@ pub enum SetExpr {
     Table(Box<Table>),
 }
 
+impl SetExpr {
+    /// If this `SetExpr` is a `SELECT`, returns the [`Select`].
+    pub fn as_select(&self) -> Option<&Select> {
+        if let Self::Select(select) = self {
+            Some(&**select)
+        } else {
+            None
+        }
+    }
+}
+
 impl fmt::Display for SetExpr {
     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
         match self {

src/parser/mod.rs

@@ -8594,8 +8594,19 @@ impl<'a> Parser<'a> {
                 self.expected("joined table", self.peek_token())
             }
         } else if dialect_of!(self is SnowflakeDialect | DatabricksDialect | GenericDialect)
-            && self.parse_keyword(Keyword::VALUES)
+            && matches!(
+                self.peek_tokens(),
+                [
+                    Token::Word(Word {
+                        keyword: Keyword::VALUES,
+                        ..
+                    }),
+                    Token::LParen
+                ]
+            )
         {
+            self.expect_keyword(Keyword::VALUES)?;
+
             // Snowflake and Databricks allow syntax like below:
             // SELECT * FROM VALUES (1, 'a'), (2, 'b') AS t (col1, col2)
             // where there are no parentheses around the VALUES clause.

tests/sqlparser_databricks.rs

@@ -1,5 +1,5 @@
 use sqlparser::ast::*;
-use sqlparser::dialect::DatabricksDialect;
+use sqlparser::dialect::{DatabricksDialect, GenericDialect};
 use sqlparser::parser::ParserError;
 use test_utils::*;
 
@@ -13,6 +13,13 @@ fn databricks() -> TestedDialects {
     }
 }
 
+fn databricks_and_generic() -> TestedDialects {
+    TestedDialects {
+        dialects: vec![Box::new(DatabricksDialect {}), Box::new(GenericDialect {})],
+        options: None,
+    }
+}
+
 #[test]
 fn test_databricks_identifiers() {
     // databricks uses backtick for delimited identifiers
@@ -124,3 +131,60 @@ fn test_databricks_lambdas() {
     );
     databricks().verified_expr("transform(array(1, 2, 3), x -> x + 1)");
 }
+
+#[test]
+fn test_values_clause() {
+    let values = Values {
+        explicit_row: false,
+        rows: vec![
+            vec![
+                Expr::Value(Value::DoubleQuotedString("one".to_owned())),
+                Expr::Value(number("1")),
+            ],
+            vec![
+                Expr::Value(Value::SingleQuotedString("two".to_owned())),
+                Expr::Value(number("2")),
+            ],
+        ],
+    };
+
+    let query = databricks().verified_query(r#"VALUES ("one", 1), ('two', 2)"#);
+    assert_eq!(SetExpr::Values(values.clone()), *query.body);
+
+    // VALUES is permitted in a FROM clause without a subquery
+    let query = databricks().verified_query_with_canonical(
+        r#"SELECT * FROM VALUES ("one", 1), ('two', 2)"#,
+        r#"SELECT * FROM (VALUES ("one", 1), ('two', 2))"#,
+    );
+    let Some(TableFactor::Derived { subquery, .. }) = query
+        .body
+        .as_select()
+        .map(|select| &select.from[0].relation)
+    else {
+        panic!("expected subquery");
+    };
+    assert_eq!(SetExpr::Values(values), *subquery.body);
+
+    // values is also a valid table name
+    let query = databricks_and_generic().verified_query(concat!(
+        "WITH values AS (SELECT 42) ",
+        "SELECT * FROM values",
+    ));
+    assert_eq!(
+        Some(&TableFactor::Table {
+            name: ObjectName(vec![Ident::new("values")]),
+            alias: None,
+            args: None,
+            with_hints: vec![],
+            version: None,
+            partitions: vec![]
+        }),
+        query
+            .body
+            .as_select()
+            .map(|select| &select.from[0].relation)
+    );
+
+    // TODO: support this example from https://docs.databricks.com/en/sql/language-manual/sql-ref-syntax-qry-select-values.html#examples
+    // databricks().verified_query("VALUES 1, 2, 3");
+}