mirrors/datafusion-sqlparse
1
0
Fork 0
mirror of https://github.com/apache/datafusion-sqlparser-rs.git synced 2025-07-07 17:04:59 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 3

Merge remote-tracking branch 'origin/main'

Browse source
This commit is contained in:
Dima 2025-06-24 14:12:26 +01:00
commit b42f58dfec
8 changed files with 293 additions and 45 deletions
Download patch file Download diff file Expand all files Collapse all files

23
src/ast/ddl.rs
View file

@ -30,11 +30,11 @@ use sqlparser_derive::{Visit, VisitMut};
use crate::ast::value::escape_single_quote_string;
use crate::ast::{
display_comma_separated, display_separated, CommentDef, CreateFunctionBody,
display_comma_separated, display_separated, ArgMode, CommentDef, CreateFunctionBody,
CreateFunctionUsing, DataType, Expr, FunctionBehavior, FunctionCalledOnNull,
FunctionDeterminismSpecifier, FunctionParallel, Ident, MySQLColumnPosition, ObjectName,
OperateFunctionArg, OrderByExpr, ProjectionSelect, SequenceOptions, SqlOption, Tag, Value,
ValueWithSpan,
FunctionDeterminismSpecifier, FunctionParallel, Ident, IndexColumn, MySQLColumnPosition,
ObjectName, OperateFunctionArg, OrderByExpr, ProjectionSelect, SequenceOptions, SqlOption, Tag,
Value, ValueWithSpan,
};
use crate::keywords::Keyword;
use crate::tokenizer::Token;
@ -979,7 +979,7 @@ pub enum TableConstraint {
/// [1]: IndexType
index_type: Option<IndexType>,
/// Identifiers of the columns that are unique.
columns: Vec<Ident>,
columns: Vec<IndexColumn>,
index_options: Vec<IndexOption>,
characteristics: Option<ConstraintCharacteristics>,
/// Optional Postgres nulls handling: `[ NULLS [ NOT ] DISTINCT ]`
@ -1015,7 +1015,7 @@ pub enum TableConstraint {
/// [1]: IndexType
index_type: Option<IndexType>,
/// Identifiers of the columns that form the primary key.
columns: Vec<Ident>,
columns: Vec<IndexColumn>,
index_options: Vec<IndexOption>,
characteristics: Option<ConstraintCharacteristics>,
},
@ -1060,7 +1060,7 @@ pub enum TableConstraint {
/// [1]: IndexType
index_type: Option<IndexType>,
/// Referred column identifier list.
columns: Vec<Ident>,
columns: Vec<IndexColumn>,
},
/// MySQLs [fulltext][1] definition. Since the [`SPATIAL`][2] definition is exactly the same,
/// and MySQL displays both the same way, it is part of this definition as well.
@ -1083,7 +1083,7 @@ pub enum TableConstraint {
/// Optional index name.
opt_index_name: Option<Ident>,
/// Referred column identifier list.
columns: Vec<Ident>,
columns: Vec<IndexColumn>,
},
}
@ -1367,11 +1367,16 @@ impl fmt::Display for NullsDistinctOption {
pub struct ProcedureParam {
pub name: Ident,
pub data_type: DataType,
pub mode: Option<ArgMode>,
}
impl fmt::Display for ProcedureParam {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
write!(f, "{} {}", self.name, self.data_type)
if let Some(mode) = &self.mode {
write!(f, "{mode} {} {}", self.name, self.data_type)
} else {
write!(f, "{} {}", self.name, self.data_type)
}
}
}

35
src/ast/spans.rs
View file

@ -28,16 +28,17 @@ use super::{
ConstraintCharacteristics, CopySource, CreateIndex, CreateTable, CreateTableOptions, Cte,
Delete, DoUpdate, ExceptSelectItem, ExcludeSelectItem, Expr, ExprWithAlias, Fetch, FromTable,
Function, FunctionArg, FunctionArgExpr, FunctionArgumentClause, FunctionArgumentList,
FunctionArguments, GroupByExpr, HavingBound, IfStatement, IlikeSelectItem, Insert, Interpolate,
InterpolateExpr, Join, JoinConstraint, JoinOperator, JsonPath, JsonPathElem, LateralView,
LimitClause, MatchRecognizePattern, Measure, NamedParenthesizedList, NamedWindowDefinition,
ObjectName, ObjectNamePart, Offset, OnConflict, OnConflictAction, OnInsert, OpenStatement,
OrderBy, OrderByExpr, OrderByKind, Partition, PivotValueSource, ProjectionSelect, Query,
RaiseStatement, RaiseStatementValue, ReferentialAction, RenameSelectItem, ReplaceSelectElement,
ReplaceSelectItem, Select, SelectInto, SelectItem, SetExpr, SqlOption, Statement, Subscript,
SymbolDefinition, TableAlias, TableAliasColumnDef, TableConstraint, TableFactor, TableObject,
TableOptionsClustered, TableWithJoins, UpdateTableFromKind, Use, Value, Values, ViewColumnDef,
WhileStatement, WildcardAdditionalOptions, With, WithFill,
FunctionArguments, GroupByExpr, HavingBound, IfStatement, IlikeSelectItem, IndexColumn, Insert,
Interpolate, InterpolateExpr, Join, JoinConstraint, JoinOperator, JsonPath, JsonPathElem,
LateralView, LimitClause, MatchRecognizePattern, Measure, NamedParenthesizedList,
NamedWindowDefinition, ObjectName, ObjectNamePart, Offset, OnConflict, OnConflictAction,
OnInsert, OpenStatement, OrderBy, OrderByExpr, OrderByKind, Partition, PivotValueSource,
ProjectionSelect, Query, RaiseStatement, RaiseStatementValue, ReferentialAction,
RenameSelectItem, ReplaceSelectElement, ReplaceSelectItem, Select, SelectInto, SelectItem,
SetExpr, SqlOption, Statement, Subscript, SymbolDefinition, TableAlias, TableAliasColumnDef,
TableConstraint, TableFactor, TableObject, TableOptionsClustered, TableWithJoins,
UpdateTableFromKind, Use, Value, Values, ViewColumnDef, WhileStatement,
WildcardAdditionalOptions, With, WithFill,
};
/// Given an iterator of spans, return the [Span::union] of all spans.
@ -650,7 +651,7 @@ impl Spanned for TableConstraint {
name.iter()
.map(|i| i.span)
.chain(index_name.iter().map(|i| i.span))
.chain(columns.iter().map(|i| i.span))
.chain(columns.iter().map(|i| i.span()))
.chain(characteristics.iter().map(|i| i.span())),
),
TableConstraint::PrimaryKey {
@ -664,7 +665,7 @@ impl Spanned for TableConstraint {
name.iter()
.map(|i| i.span)
.chain(index_name.iter().map(|i| i.span))
.chain(columns.iter().map(|i| i.span))
.chain(columns.iter().map(|i| i.span()))
.chain(characteristics.iter().map(|i| i.span())),
),
TableConstraint::ForeignKey {
@ -700,7 +701,7 @@ impl Spanned for TableConstraint {
} => union_spans(
name.iter()
.map(|i| i.span)
.chain(columns.iter().map(|i| i.span)),
.chain(columns.iter().map(|i| i.span())),
),
TableConstraint::FulltextOrSpatial {
fulltext: _,
@ -711,7 +712,7 @@ impl Spanned for TableConstraint {
opt_index_name
.iter()
.map(|i| i.span)
.chain(columns.iter().map(|i| i.span)),
.chain(columns.iter().map(|i| i.span())),
),
}
}
@ -745,6 +746,12 @@ impl Spanned for CreateIndex {
}
}
impl Spanned for IndexColumn {
fn span(&self) -> Span {
self.column.span()
}
}
impl Spanned for CaseStatement {
fn span(&self) -> Span {
let CaseStatement {

74
src/parser/mod.rs
View file

@ -6868,9 +6868,7 @@ impl<'a> Parser<'a> {
None
};
self.expect_token(&Token::LParen)?;
let columns = self.parse_comma_separated(Parser::parse_create_index_expr)?;
self.expect_token(&Token::RParen)?;
let columns = self.parse_parenthesized_index_column_list()?;
let include = if self.parse_keyword(Keyword::INCLUDE) {
self.expect_token(&Token::LParen)?;
@ -7626,9 +7624,22 @@ impl<'a> Parser<'a> {
}
pub fn parse_procedure_param(&mut self) -> Result<ProcedureParam, ParserError> {
let mode = if self.parse_keyword(Keyword::IN) {
Some(ArgMode::In)
} else if self.parse_keyword(Keyword::OUT) {
Some(ArgMode::Out)
} else if self.parse_keyword(Keyword::INOUT) {
Some(ArgMode::InOut)
} else {
None
};
let name = self.parse_identifier()?;
let data_type = self.parse_data_type()?;
Ok(ProcedureParam { name, data_type })
Ok(ProcedureParam {
name,
data_type,
mode,
})
}
pub fn parse_column_def(&mut self) -> Result<ColumnDef, ParserError> {
@ -8070,7 +8081,7 @@ impl<'a> Parser<'a> {
let index_name = self.parse_optional_ident()?;
let index_type = self.parse_optional_using_then_index_type()?;
let columns = self.parse_parenthesized_column_list(Mandatory, false)?;
let columns = self.parse_parenthesized_index_column_list()?;
let index_options = self.parse_index_options()?;
let characteristics = self.parse_constraint_characteristics()?;
Ok(Some(TableConstraint::Unique {
@ -8092,7 +8103,7 @@ impl<'a> Parser<'a> {
let index_name = self.parse_optional_ident()?;
let index_type = self.parse_optional_using_then_index_type()?;
let columns = self.parse_parenthesized_column_list(Mandatory, false)?;
let columns = self.parse_parenthesized_index_column_list()?;
let index_options = self.parse_index_options()?;
let characteristics = self.parse_constraint_characteristics()?;
Ok(Some(TableConstraint::PrimaryKey {
@ -8170,7 +8181,7 @@ impl<'a> Parser<'a> {
};
let index_type = self.parse_optional_using_then_index_type()?;
let columns = self.parse_parenthesized_column_list(Mandatory, false)?;
let columns = self.parse_parenthesized_index_column_list()?;
Ok(Some(TableConstraint::Index {
display_as_key,
@ -8199,7 +8210,7 @@ impl<'a> Parser<'a> {
let opt_index_name = self.parse_optional_ident()?;
let columns = self.parse_parenthesized_column_list(Mandatory, false)?;
let columns = self.parse_parenthesized_index_column_list()?;
Ok(Some(TableConstraint::FulltextOrSpatial {
fulltext,
@ -10601,6 +10612,14 @@ impl<'a> Parser<'a> {
self.parse_parenthesized_column_list_inner(optional, allow_empty, |p| p.parse_identifier())
}
/// Parses a parenthesized comma-separated list of index columns, which can be arbitrary
/// expressions with ordering information (and an opclass in some dialects).
fn parse_parenthesized_index_column_list(&mut self) -> Result<Vec<IndexColumn>, ParserError> {
self.parse_parenthesized_column_list_inner(Mandatory, false, |p| {
p.parse_create_index_expr()
})
}
/// Parses a parenthesized comma-separated list of qualified, possibly quoted identifiers.
/// For example: `(db1.sc1.tbl1.col1, db1.sc1.tbl1."col 2", ...)`
pub fn parse_parenthesized_qualified_column_list(
@ -15017,7 +15036,8 @@ impl<'a> Parser<'a> {
/// Parse a FETCH clause
pub fn parse_fetch(&mut self) -> Result<Fetch, ParserError> {
self.expect_one_of_keywords(&[Keyword::FIRST, Keyword::NEXT])?;
let _ = self.parse_one_of_keywords(&[Keyword::FIRST, Keyword::NEXT]);
let (quantity, percent) = if self
.parse_one_of_keywords(&[Keyword::ROW, Keyword::ROWS])
.is_some()
@ -15026,16 +15046,16 @@ impl<'a> Parser<'a> {
} else {
let quantity = Expr::Value(self.parse_value()?);
let percent = self.parse_keyword(Keyword::PERCENT);
self.expect_one_of_keywords(&[Keyword::ROW, Keyword::ROWS])?;
let _ = self.parse_one_of_keywords(&[Keyword::ROW, Keyword::ROWS]);
(Some(quantity), percent)
};
let with_ties = if self.parse_keyword(Keyword::ONLY) {
false
} else if self.parse_keywords(&[Keyword::WITH, Keyword::TIES]) {
true
} else {
return self.expected("one of ONLY or WITH TIES", self.peek_token());
self.parse_keywords(&[Keyword::WITH, Keyword::TIES])
};
Ok(Fetch {
with_ties,
percent,
@ -16527,6 +16547,20 @@ mod tests {
}};
}
fn mk_expected_col(name: &str) -> IndexColumn {
IndexColumn {
column: OrderByExpr {
expr: Expr::Identifier(name.into()),
options: OrderByOptions {
asc: None,
nulls_first: None,
},
with_fill: None,
},
operator_class: None,
}
}
let dialect =
TestedDialects::new(vec![Box::new(GenericDialect {}), Box::new(MySqlDialect {})]);
@ -16537,7 +16571,7 @@ mod tests {
display_as_key: false,
name: None,
index_type: None,
columns: vec![Ident::new("c1")],
columns: vec![mk_expected_col("c1")],
}
);
@ -16548,7 +16582,7 @@ mod tests {
display_as_key: true,
name: None,
index_type: None,
columns: vec![Ident::new("c1")],
columns: vec![mk_expected_col("c1")],
}
);
@ -16559,7 +16593,7 @@ mod tests {
display_as_key: false,
name: Some(Ident::with_quote('\'', "index")),
index_type: None,
columns: vec![Ident::new("c1"), Ident::new("c2")],
columns: vec![mk_expected_col("c1"), mk_expected_col("c2")],
}
);
@ -16570,7 +16604,7 @@ mod tests {
display_as_key: false,
name: None,
index_type: Some(IndexType::BTree),
columns: vec![Ident::new("c1")],
columns: vec![mk_expected_col("c1")],
}
);
@ -16581,7 +16615,7 @@ mod tests {
display_as_key: false,
name: None,
index_type: Some(IndexType::Hash),
columns: vec![Ident::new("c1")],
columns: vec![mk_expected_col("c1")],
}
);
@ -16592,7 +16626,7 @@ mod tests {
display_as_key: false,
name: Some(Ident::new("idx_name")),
index_type: Some(IndexType::BTree),
columns: vec![Ident::new("c1")],
columns: vec![mk_expected_col("c1")],
}
);
@ -16603,7 +16637,7 @@ mod tests {
display_as_key: false,
name: Some(Ident::new("idx_name")),
index_type: Some(IndexType::Hash),
columns: vec![Ident::new("c1")],
columns: vec![mk_expected_col("c1")],
}
);
}

44
src/test_utils.rs
View file

@ -448,3 +448,47 @@ pub fn call(function: &str, args: impl IntoIterator<Item = Expr>) -> Expr {
within_group: vec![],
})
}
/// Gets the first index column (mysql calls it a key part) of the first index found in a
/// [`Statement::CreateIndex`], [`Statement::CreateTable`], or [`Statement::AlterTable`].
pub fn index_column(stmt: Statement) -> Expr {
match stmt {
Statement::CreateIndex(CreateIndex { columns, .. }) => {
columns.first().unwrap().column.expr.clone()
}
Statement::CreateTable(CreateTable { constraints, .. }) => {
match constraints.first().unwrap() {
TableConstraint::Index { columns, .. } => {
columns.first().unwrap().column.expr.clone()
}
TableConstraint::Unique { columns, .. } => {
columns.first().unwrap().column.expr.clone()
}
TableConstraint::PrimaryKey { columns, .. } => {
columns.first().unwrap().column.expr.clone()
}
TableConstraint::FulltextOrSpatial { columns, .. } => {
columns.first().unwrap().column.expr.clone()
}
_ => panic!("Expected an index, unique, primary, full text, or spatial constraint (foreign key does not support general key part expressions)"),
}
}
Statement::AlterTable { operations, .. } => match operations.first().unwrap() {
AlterTableOperation::AddConstraint(TableConstraint::Index { columns, .. }) => {
columns.first().unwrap().column.expr.clone()
}
AlterTableOperation::AddConstraint(TableConstraint::Unique { columns, .. }) => {
columns.first().unwrap().column.expr.clone()
}
AlterTableOperation::AddConstraint(TableConstraint::PrimaryKey { columns, .. }) => {
columns.first().unwrap().column.expr.clone()
}
AlterTableOperation::AddConstraint(TableConstraint::FulltextOrSpatial {
columns,
..
}) => columns.first().unwrap().column.expr.clone(),
_ => panic!("Expected an index, unique, primary, full text, or spatial constraint (foreign key does not support general key part expressions)"),
},
_ => panic!("Expected CREATE INDEX, ALTER TABLE, or CREATE TABLE, got: {stmt:?}"),
}
}

62
tests/sqlparser_common.rs
View file

@ -15378,3 +15378,65 @@ fn join_precedence() {
"SELECT * FROM t1 NATURAL JOIN (t5 INNER JOIN t0 ON (t0.v1 + t5.v0) > 0) WHERE t0.v1 = t1.v0",
);
}
#[test]
fn parse_create_procedure_with_parameter_modes() {
let sql = r#"CREATE PROCEDURE test_proc (IN a INTEGER, OUT b TEXT, INOUT c TIMESTAMP, d BOOL) AS BEGIN SELECT 1; END"#;
match verified_stmt(sql) {
Statement::CreateProcedure {
or_alter,
name,
params,
..
} => {
assert_eq!(or_alter, false);
assert_eq!(name.to_string(), "test_proc");
let fake_span = Span {
start: Location { line: 0, column: 0 },
end: Location { line: 0, column: 0 },
};
assert_eq!(
params,
Some(vec![
ProcedureParam {
name: Ident {
value: "a".into(),
quote_style: None,
span: fake_span,
},
data_type: DataType::Integer(None),
mode: Some(ArgMode::In)
},
ProcedureParam {
name: Ident {
value: "b".into(),
quote_style: None,
span: fake_span,
},
data_type: DataType::Text,
mode: Some(ArgMode::Out)
},
ProcedureParam {
name: Ident {
value: "c".into(),
quote_style: None,
span: fake_span,
},
data_type: DataType::Timestamp(None, TimezoneInfo::None),
mode: Some(ArgMode::InOut)
},
ProcedureParam {
name: Ident {
value: "d".into(),
quote_style: None,
span: fake_span,
},
data_type: DataType::Bool,
mode: None
},
])
);
}
_ => unreachable!(),
}
}

6
tests/sqlparser_mssql.rs
View file

@ -153,7 +153,8 @@ fn parse_create_procedure() {
quote_style: None,
span: Span::empty(),
},
data_type: DataType::Int(None)
data_type: DataType::Int(None),
mode: None,
},
ProcedureParam {
name: Ident {
@ -164,7 +165,8 @@ fn parse_create_procedure() {
data_type: DataType::Varchar(Some(CharacterLength::IntegerLength {
length: 256,
unit: None
}))
})),
mode: None,
}
]),
name: ObjectName::from(vec![Ident {

75
tests/sqlparser_mysql.rs
View file

@ -670,6 +670,20 @@ fn table_constraint_unique_primary_ctor(
characteristics: Option<ConstraintCharacteristics>,
unique_index_type_display: Option<KeyOrIndexDisplay>,
) -> TableConstraint {
let columns = columns
.into_iter()
.map(|ident| IndexColumn {
column: OrderByExpr {
expr: Expr::Identifier(ident),
options: OrderByOptions {
asc: None,
nulls_first: None,
},
with_fill: None,
},
operator_class: None,
})
.collect();
match unique_index_type_display {
Some(index_type_display) => TableConstraint::Unique {
name,
@ -795,6 +809,67 @@ fn parse_create_table_primary_and_unique_key_with_index_options() {
}
}
#[test]
fn parse_prefix_key_part() {
let expected = vec![FunctionArg::Unnamed(FunctionArgExpr::Expr(Expr::value(
number("10"),
)))];
for sql in [
"CREATE INDEX idx_index ON t(textcol(10))",
"ALTER TABLE tab ADD INDEX idx_index (textcol(10))",
"ALTER TABLE tab ADD PRIMARY KEY (textcol(10))",
"ALTER TABLE tab ADD UNIQUE KEY (textcol(10))",
"ALTER TABLE tab ADD UNIQUE KEY (textcol(10))",
"ALTER TABLE tab ADD FULLTEXT INDEX (textcol(10))",
"CREATE TABLE t (textcol TEXT, INDEX idx_index (textcol(10)))",
] {
match index_column(mysql_and_generic().verified_stmt(sql)) {
Expr::Function(Function {
name,
args: FunctionArguments::List(FunctionArgumentList { args, .. }),
..
}) => {
assert_eq!(name.to_string(), "textcol");
assert_eq!(args, expected);
}
expr => panic!("unexpected expression {expr} for {sql}"),
}
}
}
#[test]
fn test_functional_key_part() {
assert_eq!(
index_column(
mysql_and_generic()
.verified_stmt("CREATE INDEX idx_index ON t((col COLLATE utf8mb4_bin) DESC)")
),
Expr::Nested(Box::new(Expr::Collate {
expr: Box::new(Expr::Identifier("col".into())),
collation: ObjectName(vec![sqlparser::ast::ObjectNamePart::Identifier(
Ident::new("utf8mb4_bin")
)]),
}))
);
assert_eq!(
index_column(mysql_and_generic().verified_stmt(
r#"CREATE TABLE t (jsoncol JSON, PRIMARY KEY ((CAST(col ->> '$.id' AS UNSIGNED)) ASC))"#
)),
Expr::Nested(Box::new(Expr::Cast {
kind: CastKind::Cast,
expr: Box::new(Expr::BinaryOp {
left: Box::new(Expr::Identifier(Ident::new("col"))),
op: BinaryOperator::LongArrow,
right: Box::new(Expr::Value(
Value::SingleQuotedString("$.id".to_string()).with_empty_span()
)),
}),
data_type: DataType::Unsigned,
format: None,
})),
);
}
#[test]
fn parse_create_table_primary_and_unique_key_with_index_type() {
let sqls = ["UNIQUE", "PRIMARY KEY"].map(|key_ty| {

19
tests/sqlparser_snowflake.rs
View file

@ -4146,3 +4146,22 @@ END
assert_eq!(2, exception[1].idents.len());
assert_eq!(2, exception[1].statements.len());
}
#[test]
fn test_snowflake_fetch_clause_syntax() {
let canonical = "SELECT c1 FROM fetch_test FETCH FIRST 2 ROWS ONLY";
snowflake().verified_only_select_with_canonical("SELECT c1 FROM fetch_test FETCH 2", canonical);
snowflake()
.verified_only_select_with_canonical("SELECT c1 FROM fetch_test FETCH FIRST 2", canonical);
snowflake()
.verified_only_select_with_canonical("SELECT c1 FROM fetch_test FETCH NEXT 2", canonical);
snowflake()
.verified_only_select_with_canonical("SELECT c1 FROM fetch_test FETCH 2 ROW", canonical);
snowflake().verified_only_select_with_canonical(
"SELECT c1 FROM fetch_test FETCH FIRST 2 ROWS",
canonical,
);
}