Merge remote-tracking branch 'apache/main' into reduce-token-cloning

Author: alamb

Cargo.toml

@@ -37,15 +37,18 @@ name = "sqlparser"
 path = "src/lib.rs"
 
 [features]
-default = ["std"]
+default = ["std", "recursive-protection"]
 std = []
+recursive-protection = ["std", "recursive"]
 # Enable JSON output in the `cli` example:
 json_example = ["serde_json", "serde"]
 visitor = ["sqlparser_derive"]
 
 [dependencies]
 bigdecimal = { version = "0.4.1", features = ["serde"], optional = true }
 log = "0.4"
+recursive = { version = "0.1.1", optional = true}
+
 serde = { version = "1.0", features = ["derive"], optional = true }
 # serde_json is only used in examples/cli, but we have to put it outside
 # of dev-dependencies because of

README.md

@@ -63,7 +63,7 @@ The following optional [crate  features](https://doc.rust-lang.org/cargo/referen
 
 * `serde`: Adds [Serde](https://serde.rs/) support by implementing  `Serialize` and `Deserialize` for all AST nodes.
 * `visitor`: Adds a `Visitor` capable of recursively walking the AST tree.
-
+* `recursive-protection` (enabled by default), uses [recursive](https://docs.rs/recursive/latest/recursive/) for stack overflow protection. 
 
 ## Syntax vs Semantics
 

derive/src/lib.rs

@@ -78,7 +78,10 @@ fn derive_visit(input: proc_macro::TokenStream, visit_type: &VisitType) -> proc_
 
     let expanded = quote! {
         // The generated impl.
+        // Note that it uses [`recursive::recursive`] to protect from stack overflow.
+        // See tests in https://github.com/apache/datafusion-sqlparser-rs/pull/1522/ for more info.
         impl #impl_generics sqlparser::ast::#visit_trait for #name #ty_generics #where_clause {
+             #[cfg_attr(feature = "recursive-protection", recursive::recursive)]
             fn visit<V: sqlparser::ast::#visitor_trait>(
                 &#modifier self,
                 visitor: &mut V

sqlparser_bench/benches/sqlparser_bench.rs

@@ -42,6 +42,46 @@ fn basic_queries(c: &mut Criterion) {
     group.bench_function("sqlparser::with_select", |b| {
         b.iter(|| Parser::parse_sql(&dialect, with_query).unwrap());
     });
+
+    let large_statement = {
+        let expressions = (0..1000)
+            .map(|n| format!("FN_{}(COL_{})", n, n))
+            .collect::<Vec<_>>()
+            .join(", ");
+        let tables = (0..1000)
+            .map(|n| format!("TABLE_{}", n))
+            .collect::<Vec<_>>()
+            .join(" JOIN ");
+        let where_condition = (0..1000)
+            .map(|n| format!("COL_{} = {}", n, n))
+            .collect::<Vec<_>>()
+            .join(" OR ");
+        let order_condition = (0..1000)
+            .map(|n| format!("COL_{} DESC", n))
+            .collect::<Vec<_>>()
+            .join(", ");
+
+        format!(
+            "SELECT {} FROM {} WHERE {} ORDER BY {}",
+            expressions, tables, where_condition, order_condition
+        )
+    };
+
+    group.bench_function("parse_large_statement", |b| {
+        b.iter(|| Parser::parse_sql(&dialect, criterion::black_box(large_statement.as_str())));
+    });
+
+    let large_statement = Parser::parse_sql(&dialect, large_statement.as_str())
+        .unwrap()
+        .pop()
+        .unwrap();
+
+    group.bench_function("format_large_statement", |b| {
+        b.iter(|| {
+            let formatted_query = large_statement.to_string();
+            assert_eq!(formatted_query, large_statement);
+        });
+    });
 }
 
 criterion_group!(benches, basic_queries);

src/ast/data_type.rs

@@ -254,7 +254,7 @@ pub enum DataType {
     /// [postgresql]: https://www.postgresql.org/docs/15/datatype.html
     Float8,
     /// Double
-    Double,
+    Double(ExactNumberInfo),
     /// Double PRECISION e.g. [standard], [postgresql]
     ///
     /// [standard]: https://jakewheat.github.io/sql-overview/sql-2016-foundation-grammar.html#approximate-numeric-type
@@ -373,6 +373,10 @@ pub enum DataType {
     ///
     /// [postgresql]: https://www.postgresql.org/docs/current/plpgsql-trigger.html
     Trigger,
+    /// Any data type, used in BigQuery UDF definitions for templated parameters
+    ///
+    /// [bigquery]: https://cloud.google.com/bigquery/docs/user-defined-functions#templated-sql-udf-parameters
+    AnyType,
 }
 
 impl fmt::Display for DataType {
@@ -383,7 +387,6 @@ impl fmt::Display for DataType {
             DataType::CharacterVarying(size) => {
                 format_character_string_type(f, "CHARACTER VARYING", size)
             }
-
             DataType::CharVarying(size) => format_character_string_type(f, "CHAR VARYING", size),
             DataType::Varchar(size) => format_character_string_type(f, "VARCHAR", size),
             DataType::Nvarchar(size) => format_character_string_type(f, "NVARCHAR", size),
@@ -505,7 +508,7 @@ impl fmt::Display for DataType {
             DataType::Float4 => write!(f, "FLOAT4"),
             DataType::Float32 => write!(f, "Float32"),
             DataType::Float64 => write!(f, "FLOAT64"),
-            DataType::Double => write!(f, "DOUBLE"),
+            DataType::Double(info) => write!(f, "DOUBLE{info}"),
             DataType::Float8 => write!(f, "FLOAT8"),
             DataType::DoublePrecision => write!(f, "DOUBLE PRECISION"),
             DataType::Bool => write!(f, "BOOL"),
@@ -626,6 +629,7 @@ impl fmt::Display for DataType {
             }
             DataType::Unspecified => Ok(()),
             DataType::Trigger => write!(f, "TRIGGER"),
+            DataType::AnyType => write!(f, "ANY TYPE"),
         }
     }
 }

src/ast/ddl.rs

@@ -885,12 +885,14 @@ impl fmt::Display for TableConstraint {
             } => {
                 write!(
                     f,
-                    "{}FOREIGN KEY ({}) REFERENCES {}({})",
+                    "{}FOREIGN KEY ({}) REFERENCES {}",
                     display_constraint_name(name),
                     display_comma_separated(columns),
                     foreign_table,
-                    display_comma_separated(referred_columns),
                 )?;
+                if !referred_columns.is_empty() {
+                    write!(f, "({})", display_comma_separated(referred_columns))?;
+                }
                 if let Some(action) = on_delete {
                     write!(f, " ON DELETE {action}")?;
                 }