[snowflake] Support specifying an alias after FROM (table_factor)

Snowflake diverges from the standard and from most of the other
implementations by allowing extra parentheses not only around a join,
but around lone table names (e.g. `FROM (mytable [AS alias])`) and
around derived tables (e.g. `FROM ((SELECT ...)  [AS alias])`) as well.

Initially this was implemented in apache#154
by (ab)using `TableFactor::NestedJoin` to represent anything nested in
extra set of parens.

Afterwards we learned in apache#223
that in cases of such extraneous nesting Snowflake allows specifying the
alias both inside and outside parens, but not both - consider:

    FROM (table_factor AS inner_alias) AS outer_alias

We've considered implementing this by changing `TableFactor::NestedJoin`
to a `TableFactor::Nested { inner: TableWithJoins, alias:
Option<TableAlias> }`, but that seemed too generic, as no known dialect
supports duplicate aliases, as shown above, nor naming nested joins
`(foo NATURAL JOIN bar) alias`. So we decided on making a smaller change
(with no modifications to the AST), that is also more appropriate to the
contributors to the Snowflake dialect:


1) Revert apache#154 by rejecting `FROM (table or derived table)` in most dialects.

2) For `dialect_of!(self is SnowflakeDialect | GenericDialect)` parse
and strip the extraneous parentheses, e.g.

   `(mytable) AS alias` -> `(mytable AS alias)`


Co-authored-by: Eyal Leshem <[email protected]>

Author: nickolay

src/ast/query.rs

@@ -242,9 +242,10 @@ pub enum TableFactor {
     },
     /// Represents a parenthesized table factor. The SQL spec only allows a
     /// join expression (`(foo <JOIN> bar [ <JOIN> baz ... ])`) to be nested,
-    /// possibly several times, but the parser also accepts the non-standard
-    /// nesting of bare tables (`table_with_joins.joins.is_empty()`), so the
-    /// name `NestedJoin` is a bit of misnomer.
+    /// possibly several times.
+    ///
+    /// The parser may also accept non-standard nesting of bare tables for some
+    /// dialects, but the information about such nesting is stripped from AST.
     NestedJoin(Box<TableWithJoins>),
 }
 

src/parser.rs

@@ -2156,14 +2156,58 @@ impl<'a> Parser<'a> {
             // recently consumed does not start a derived table (cases 1, 2, or 4).
             // `maybe_parse` will ignore such an error and rewind to be after the opening '('.
 
-            // Inside the parentheses we expect to find a table factor
-            // followed by some joins or another level of nesting.
-            let table_and_joins = self.parse_table_and_joins()?;
-            self.expect_token(&Token::RParen)?;
-            // The SQL spec prohibits derived and bare tables from appearing
-            // alone in parentheses. We don't enforce this as some databases
-            // (e.g. Snowflake) allow such syntax.
-            Ok(TableFactor::NestedJoin(Box::new(table_and_joins)))
+            // Inside the parentheses we expect to find an (A) table factor
+            // followed by some joins or (B) another level of nesting.
+            let mut table_and_joins = self.parse_table_and_joins()?;
+
+            if !table_and_joins.joins.is_empty() {
+                self.expect_token(&Token::RParen)?;
+                Ok(TableFactor::NestedJoin(Box::new(table_and_joins))) // (A)
+            } else if let TableFactor::NestedJoin(_) = &table_and_joins.relation {
+                // (B): `table_and_joins` (what we found inside the parentheses)
+                // is a nested join `(foo JOIN bar)`, not followed by other joins.
+                self.expect_token(&Token::RParen)?;
+                Ok(TableFactor::NestedJoin(Box::new(table_and_joins)))
+            } else if dialect_of!(self is SnowflakeDialect | GenericDialect) {
+                // Dialect-specific behavior: Snowflake diverges from the
+                // standard and from most of the other implementations by
+                // allowing extra parentheses not only around a join (B), but
+                // around lone table names (e.g. `FROM (mytable [AS alias])`)
+                // and around derived tables (e.g. `FROM ((SELECT ...)
+                // [AS alias])`) as well.
+                self.expect_token(&Token::RParen)?;
+
+                if let Some(outer_alias) =
+                    self.parse_optional_table_alias(keywords::RESERVED_FOR_TABLE_ALIAS)?
+                {
+                    // Snowflake also allows specifying an alias *after* parens
+                    // e.g. `FROM (mytable) AS alias`
+                    match &mut table_and_joins.relation {
+                        TableFactor::Derived { alias, .. }
+                        | TableFactor::Table { alias, .. }
+                        | TableFactor::TableFunction { alias, .. } => {
+                            // but not `FROM (mytable AS alias1) AS alias2`.
+                            if let Some(inner_alias) = alias {
+                                return Err(ParserError::ParserError(format!(
+                                    "duplicate alias {}",
+                                    inner_alias
+                                )));
+                            }
+                            // Act as if the alias was specified normally next
+                            // to the table name: `(mytable) AS alias` ->
+                            // `(mytable AS alias)`
+                            alias.replace(outer_alias);
+                        }
+                        TableFactor::NestedJoin(_) => unreachable!(),
+                    };
+                }
+                // Do not store the extra set of parens in the AST
+                Ok(table_and_joins.relation)
+            } else {
+                // The SQL spec prohibits derived tables and bare tables from
+                // appearing alone in parentheses (e.g. `FROM (mytable)`)
+                self.expected("joined table", self.peek_token())
+            }
         } else {
             let name = self.parse_object_name()?;
             // Postgres, MSSQL: table-valued functions:

src/test_utils.rs

@@ -156,6 +156,13 @@ pub fn number(n: &'static str) -> Value {
     Value::Number(n.parse().unwrap())
 }
 
+pub fn table_alias(name: impl Into<String>) -> Option<TableAlias> {
+    Some(TableAlias {
+        name: Ident::new(name),
+        columns: vec![],
+    })
+}
+
 pub fn table(name: impl Into<String>) -> TableFactor {
     TableFactor::Table {
         name: ObjectName(vec![Ident::new(name.into())]),

tests/sqlparser_common.rs

@@ -20,7 +20,7 @@
 
 #[macro_use]
 mod test_utils;
-use test_utils::{all_dialects, expr_from_projection, join, number, only, table};
+use test_utils::{all_dialects, expr_from_projection, join, number, only, table, table_alias};
 
 use matches::assert_matches;
 use sqlparser::ast::*;
@@ -2130,13 +2130,6 @@ fn parse_cross_join() {
     );
 }
 
-fn table_alias(name: impl Into<String>) -> Option<TableAlias> {
-    Some(TableAlias {
-        name: Ident::new(name),
-        columns: vec![],
-    })
-}
-
 #[test]
 fn parse_joins_on() {
     fn join_with_constraint(

tests/sqlparser_snowflake.rs

@@ -20,6 +20,7 @@ use test_utils::*;
 
 use sqlparser::ast::*;
 use sqlparser::dialect::{GenericDialect, SnowflakeDialect};
+use sqlparser::parser::ParserError;
 use sqlparser::tokenizer::*;
 
 #[test]
@@ -72,42 +73,81 @@ fn test_snowflake_single_line_tokenize() {
 }
 
 #[test]
-fn test_sf_derives_single_table_in_parenthesis() {
-    // Nesting a subquery in parentheses is non-standard, but supported in Snowflake SQL
-    let sql = "SELECT * FROM ((SELECT 1) AS t)";
-    let select = snowflake_and_generic().verified_only_select(sql);
-    let from = only(select.from);
-    assert_eq!(
-        from.relation,
-        TableFactor::NestedJoin(Box::new(TableWithJoins {
-            relation: TableFactor::Derived {
-                lateral: false,
-                subquery: Box::new(snowflake_and_generic().verified_query("SELECT 1")),
-                alias: Some(TableAlias {
-                    name: "t".into(),
-                    columns: vec![],
-                })
-            },
-            joins: Vec::new(),
-        }))
+fn test_sf_derived_table_in_parenthesis() {
+    // Nesting a subquery in an extra set of parentheses is non-standard,
+    // but supported in Snowflake SQL
+    snowflake_and_generic().one_statement_parses_to(
+        "SELECT * FROM ((SELECT 1) AS t)",
+        "SELECT * FROM (SELECT 1) AS t",
+    );
+    snowflake_and_generic().one_statement_parses_to(
+        "SELECT * FROM (((SELECT 1) AS t))",
+        "SELECT * FROM (SELECT 1) AS t",
     );
 }
 
 #[test]
 fn test_single_table_in_parenthesis() {
     // Parenthesized table names are non-standard, but supported in Snowflake SQL
-    let sql = "SELECT * FROM (a NATURAL JOIN (b))";
-    let select = snowflake_and_generic().verified_only_select(sql);
-    let from = only(select.from);
+    snowflake_and_generic().one_statement_parses_to(
+        "SELECT * FROM (a NATURAL JOIN (b))",
+        "SELECT * FROM (a NATURAL JOIN b)",
+    );
+    snowflake_and_generic().one_statement_parses_to(
+        "SELECT * FROM (a NATURAL JOIN ((b)))",
+        "SELECT * FROM (a NATURAL JOIN b)",
+    );
+}
 
-    assert_eq!(from.relation, nest!(table("a"), nest!(table("b"))));
+#[test]
+fn test_single_table_in_parenthesis_with_alias() {
+    snowflake_and_generic().one_statement_parses_to(
+        "SELECT * FROM (a NATURAL JOIN (b) c )",
+        "SELECT * FROM (a NATURAL JOIN b AS c)",
+    );
 
-    // Double parentheses around table names are non-standard, but supported in Snowflake SQL
-    let sql = "SELECT * FROM (a NATURAL JOIN ((b)))";
-    let select = snowflake_and_generic().verified_only_select(sql);
-    let from = only(select.from);
+    snowflake_and_generic().one_statement_parses_to(
+        "SELECT * FROM (a NATURAL JOIN ((b)) c )",
+        "SELECT * FROM (a NATURAL JOIN b AS c)",
+    );
+
+    snowflake_and_generic().one_statement_parses_to(
+        "SELECT * FROM (a NATURAL JOIN ( (b) c ) )",
+        "SELECT * FROM (a NATURAL JOIN b AS c)",
+    );
 
-    assert_eq!(from.relation, nest!(table("a"), nest!(nest!(table("b")))));
+    snowflake_and_generic().one_statement_parses_to(
+        "SELECT * FROM (a NATURAL JOIN ( (b) as c ) )",
+        "SELECT * FROM (a NATURAL JOIN b AS c)",
+    );
+
+    snowflake_and_generic().one_statement_parses_to(
+        "SELECT * FROM (a alias1 NATURAL JOIN ( (b) c ) )",
+        "SELECT * FROM (a AS alias1 NATURAL JOIN b AS c)",
+    );
+
+    snowflake_and_generic().one_statement_parses_to(
+        "SELECT * FROM (a as alias1 NATURAL JOIN ( (b) as c ) )",
+        "SELECT * FROM (a AS alias1 NATURAL JOIN b AS c)",
+    );
+
+    let res = snowflake_and_generic().parse_sql_statements("SELECT * FROM (a NATURAL JOIN b) c");
+    assert_eq!(
+        ParserError::ParserError("Expected end of statement, found: c".to_string()),
+        res.unwrap_err()
+    );
+
+    let res = snowflake().parse_sql_statements("SELECT * FROM (a b) c");
+    assert_eq!(
+        ParserError::ParserError("duplicate alias b".to_string()),
+        res.unwrap_err()
+    );
+}
+
+fn snowflake() -> TestedDialects {
+    TestedDialects {
+        dialects: vec![Box::new(SnowflakeDialect {})],
+    }
 }
 
 fn snowflake_and_generic() -> TestedDialects {