You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@datafusion.apache.org by gi...@apache.org on 2026/04/09 09:57:26 UTC
(datafusion-sqlparser-rs) branch gh-readonly-queue/main/pr-2286-ea9b413c9979badc71559ae6896c6e7aa74a696b created (now 72b295a6)
This is an automated email from the ASF dual-hosted git repository.
github-bot pushed a change to branch gh-readonly-queue/main/pr-2286-ea9b413c9979badc71559ae6896c6e7aa74a696b
in repository https://gitbox.apache.org/repos/asf/datafusion-sqlparser-rs.git
at 72b295a6 Support optional AS keyword in CTE definitions for Databricks (#2286)
This branch includes the following new commits:
new 72b295a6 Support optional AS keyword in CTE definitions for Databricks (#2286)
The 1 revisions listed above as "new" are entirely new to this
repository and will be described in separate emails. The revisions
listed as "add" were already present in the repository and have only
been added to this reference.
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@datafusion.apache.org
For additional commands, e-mail: commits-help@datafusion.apache.org
(datafusion-sqlparser-rs) 01/01: Support optional AS keyword in CTE definitions for Databricks (#2286)
Posted by gi...@apache.org.
This is an automated email from the ASF dual-hosted git repository.
github-bot pushed a commit to branch gh-readonly-queue/main/pr-2286-ea9b413c9979badc71559ae6896c6e7aa74a696b
in repository https://gitbox.apache.org/repos/asf/datafusion-sqlparser-rs.git
commit 72b295a6268eb4b3718a8cc887501e2b31ac1552
Author: Minjun Kim <48...@users.noreply.github.com>
AuthorDate: Thu Apr 9 18:57:01 2026 +0900
Support optional AS keyword in CTE definitions for Databricks (#2286)
---
src/dialect/databricks.rs | 5 +++
src/dialect/generic.rs | 4 ++
src/dialect/mod.rs | 11 +++++
src/parser/mod.rs | 98 ++++++++++++++++++++++++-------------------
tests/sqlparser_databricks.rs | 27 ++++++++++++
5 files changed, 101 insertions(+), 44 deletions(-)
diff --git a/src/dialect/databricks.rs b/src/dialect/databricks.rs
index 09cac96f..c76b464a 100644
--- a/src/dialect/databricks.rs
+++ b/src/dialect/databricks.rs
@@ -99,4 +99,9 @@ impl Dialect for DatabricksDialect {
fn supports_bang_not_operator(&self) -> bool {
true
}
+
+ /// See <https://docs.databricks.com/aws/en/sql/language-manual/sql-ref-syntax-qry-select-cte>
+ fn supports_cte_without_as(&self) -> bool {
+ true
+ }
}
diff --git a/src/dialect/generic.rs b/src/dialect/generic.rs
index 1d5461fe..c7f17351 100644
--- a/src/dialect/generic.rs
+++ b/src/dialect/generic.rs
@@ -288,4 +288,8 @@ impl Dialect for GenericDialect {
fn supports_comma_separated_trim(&self) -> bool {
true
}
+
+ fn supports_cte_without_as(&self) -> bool {
+ true
+ }
}
diff --git a/src/dialect/mod.rs b/src/dialect/mod.rs
index bbf7d580..ef9cafed 100644
--- a/src/dialect/mod.rs
+++ b/src/dialect/mod.rs
@@ -1670,6 +1670,17 @@ pub trait Dialect: Debug + Any {
fn supports_comma_separated_trim(&self) -> bool {
false
}
+
+ /// Returns true if the dialect supports the `AS` keyword being
+ /// optional in a CTE definition. For example:
+ /// ```sql
+ /// WITH cte_name (SELECT ...)
+ /// ```
+ ///
+ /// [Databricks](https://docs.databricks.com/aws/en/sql/language-manual/sql-ref-syntax-qry-select-cte)
+ fn supports_cte_without_as(&self) -> bool {
+ false
+ }
}
/// Operators for which precedence must be defined.
diff --git a/src/parser/mod.rs b/src/parser/mod.rs
index 70e8ce28..9056f85d 100644
--- a/src/parser/mod.rs
+++ b/src/parser/mod.rs
@@ -14060,64 +14060,74 @@ impl<'a> Parser<'a> {
})
}
- /// Parse a CTE (`alias [( col1, col2, ... )] AS (subquery)`)
+ /// Parse a CTE (`alias [( col1, col2, ... )] [AS] (subquery)`)
pub fn parse_cte(&mut self) -> Result<Cte, ParserError> {
let name = self.parse_identifier()?;
- let mut cte = if self.parse_keyword(Keyword::AS) {
- let mut is_materialized = None;
- if dialect_of!(self is PostgreSqlDialect) {
- if self.parse_keyword(Keyword::MATERIALIZED) {
- is_materialized = Some(CteAsMaterialized::Materialized);
- } else if self.parse_keywords(&[Keyword::NOT, Keyword::MATERIALIZED]) {
- is_materialized = Some(CteAsMaterialized::NotMaterialized);
+ let as_optional = self.dialect.supports_cte_without_as();
+
+ // If AS is optional, first try to parse `name (query)` directly
+ if as_optional && !self.peek_keyword(Keyword::AS) {
+ if let Some((query, closing_paren_token)) = self.maybe_parse(|p| {
+ p.expect_token(&Token::LParen)?;
+ let query = p.parse_query()?;
+ let closing_paren_token = p.expect_token(&Token::RParen)?;
+ Ok((query, closing_paren_token))
+ })? {
+ let mut cte = Cte {
+ alias: TableAlias {
+ explicit: false,
+ name,
+ columns: vec![],
+ },
+ query,
+ from: None,
+ materialized: None,
+ closing_paren_token: closing_paren_token.into(),
+ };
+ if self.parse_keyword(Keyword::FROM) {
+ cte.from = Some(self.parse_identifier()?);
}
+ return Ok(cte);
}
- self.expect_token(&Token::LParen)?;
-
- let query = self.parse_query()?;
- let closing_paren_token = self.expect_token(&Token::RParen)?;
+ }
- let alias = TableAlias {
- explicit: false,
- name,
- columns: vec![],
- };
- Cte {
- alias,
- query,
- from: None,
- materialized: is_materialized,
- closing_paren_token: closing_paren_token.into(),
- }
+ // Determine column definitions and consume AS
+ let columns = if self.parse_keyword(Keyword::AS) {
+ vec![]
} else {
let columns = self.parse_table_alias_column_defs()?;
- self.expect_keyword_is(Keyword::AS)?;
- let mut is_materialized = None;
- if dialect_of!(self is PostgreSqlDialect) {
- if self.parse_keyword(Keyword::MATERIALIZED) {
- is_materialized = Some(CteAsMaterialized::Materialized);
- } else if self.parse_keywords(&[Keyword::NOT, Keyword::MATERIALIZED]) {
- is_materialized = Some(CteAsMaterialized::NotMaterialized);
- }
+ if as_optional {
+ let _ = self.parse_keyword(Keyword::AS);
+ } else {
+ self.expect_keyword_is(Keyword::AS)?;
}
- self.expect_token(&Token::LParen)?;
+ columns
+ };
- let query = self.parse_query()?;
- let closing_paren_token = self.expect_token(&Token::RParen)?;
+ let mut is_materialized = None;
+ if dialect_of!(self is PostgreSqlDialect) {
+ if self.parse_keyword(Keyword::MATERIALIZED) {
+ is_materialized = Some(CteAsMaterialized::Materialized);
+ } else if self.parse_keywords(&[Keyword::NOT, Keyword::MATERIALIZED]) {
+ is_materialized = Some(CteAsMaterialized::NotMaterialized);
+ }
+ }
- let alias = TableAlias {
+ self.expect_token(&Token::LParen)?;
+ let query = self.parse_query()?;
+ let closing_paren_token = self.expect_token(&Token::RParen)?;
+
+ let mut cte = Cte {
+ alias: TableAlias {
explicit: false,
name,
columns,
- };
- Cte {
- alias,
- query,
- from: None,
- materialized: is_materialized,
- closing_paren_token: closing_paren_token.into(),
- }
+ },
+ query,
+ from: None,
+ materialized: is_materialized,
+ closing_paren_token: closing_paren_token.into(),
};
if self.parse_keyword(Keyword::FROM) {
cte.from = Some(self.parse_identifier()?);
diff --git a/tests/sqlparser_databricks.rs b/tests/sqlparser_databricks.rs
index e9b19fcc..dd7d2881 100644
--- a/tests/sqlparser_databricks.rs
+++ b/tests/sqlparser_databricks.rs
@@ -651,3 +651,30 @@ fn parse_numeric_prefix_identifier() {
databricks().verified_stmt("SELECT * FROM a.b.1c");
}
+
+#[test]
+fn parse_cte_without_as() {
+ databricks_and_generic().one_statement_parses_to(
+ "WITH cte (SELECT 1) SELECT * FROM cte",
+ "WITH cte AS (SELECT 1) SELECT * FROM cte",
+ );
+
+ databricks_and_generic().one_statement_parses_to(
+ "WITH a AS (SELECT 1), b (SELECT 2) SELECT * FROM a, b",
+ "WITH a AS (SELECT 1), b AS (SELECT 2) SELECT * FROM a, b",
+ );
+
+ databricks_and_generic().one_statement_parses_to(
+ "WITH cte (col1, col2) (SELECT 1, 2) SELECT * FROM cte",
+ "WITH cte (col1, col2) AS (SELECT 1, 2) SELECT * FROM cte",
+ );
+
+ databricks_and_generic().verified_query("WITH cte AS (SELECT 1) SELECT * FROM cte");
+
+ databricks_and_generic()
+ .verified_query("WITH cte (col1, col2) AS (SELECT 1, 2) SELECT * FROM cte");
+
+ assert!(all_dialects_where(|d| !d.supports_cte_without_as())
+ .parse_sql_statements("WITH cte (SELECT 1) SELECT * FROM cte")
+ .is_err());
+}
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@datafusion.apache.org
For additional commands, e-mail: commits-help@datafusion.apache.org