This is an automated email from the ASF dual-hosted git repository.
alamb pushed a commit to branch branch-53
in repository https://gitbox.apache.org/repos/asf/datafusion.git
The following commit(s) were added to refs/heads/branch-53 by this push:
new f8a03d87b0 [branch-53] feat: parse `JsonAccess` as a binary operator,
add `Operator::Colon` (#20717)
f8a03d87b0 is described below
commit f8a03d87b058776bb38001d3b1d56b0b673e57b0
Author: Samyak Sarnayak <[email protected]>
AuthorDate: Fri Mar 6 03:20:17 2026 +0530
[branch-53] feat: parse `JsonAccess` as a binary operator, add
`Operator::Colon` (#20717)
## Which issue does this PR close?
- part of #19692
<!--
We generally require a GitHub issue to be filed for all bug fixes and
enhancements and this helps us generate change logs for our releases.
You can link an issue to this PR using the GitHub syntax. For example
`Closes #123` indicates that this PR will close issue #123.
-->
- Needed in
https://github.com/datafusion-contrib/datafusion-variant/issues/26
- Backport of https://github.com/apache/datafusion/pull/20628 on v53
branch.
## Rationale for this change
<!--
Why are you proposing this change? If this is already explained clearly
in the issue then this section is not needed.
Explaining clearly why changes are proposed helps reviewers understand
your changes and offer better suggestions for fixes.
-->
- `sqlparser-rs` currently exposes the colon operator (`:`) as a special
`JsonAccess` expression. So it fails in datafusion's parsing before an
`ExprPlanner` is even invoked.
- Add `Operator::Colon`. Currently it's not used/implemented in
datafusion.
## What changes are included in this PR?
<!--
There is no need to duplicate the description in the issue here but it
is sometimes worth providing a summary of the individual changes in this
PR.
-->
- Fixes the above problem by converting `JsonAccess` to a normal binary
expr, on which the `ExprPlanner` is invoked and custom parsing can be
done.
## Are these changes tested?
<!--
We typically require tests for all PRs in order to:
1. Prevent the code from being accidentally broken by subsequent changes
2. Serve as another way to document the expected behavior of the code
If tests are not included in your PR, please explain why (for example,
are they covered by existing tests)?
-->
Added tests.
Also did a prototype of a custom `ExprPlanner` in datafusion-variant
using this to convert colon operator to `variant_get` function -
https://github.com/datafusion-contrib/datafusion-variant/pull/31
## Are there any user-facing changes?
<!--
If there are user-facing changes then we may require documentation to be
updated before approving the PR.
-->
<!--
If there are any breaking changes to public APIs, please add the `api
change` label.
-->
Add `Operator::Colon`
---
datafusion/expr-common/src/operator.rs | 17 +++++++---
datafusion/expr-common/src/type_coercion/binary.rs | 3 ++
datafusion/physical-expr/src/expressions/binary.rs | 2 +-
datafusion/sql/src/expr/binary_op.rs | 3 +-
datafusion/sql/src/expr/mod.rs | 30 ++++++++++++++++--
datafusion/sql/src/unparser/expr.rs | 1 +
datafusion/sql/tests/cases/plan_to_sql.rs | 36 ++++++++++++++++++++++
.../logical_plan/producer/expr/scalar_function.rs | 1 +
8 files changed, 85 insertions(+), 8 deletions(-)
diff --git a/datafusion/expr-common/src/operator.rs
b/datafusion/expr-common/src/operator.rs
index 33512b0c35..427069b326 100644
--- a/datafusion/expr-common/src/operator.rs
+++ b/datafusion/expr-common/src/operator.rs
@@ -140,6 +140,10 @@ pub enum Operator {
///
/// Not implemented in DataFusion yet.
QuestionPipe,
+ /// Colon operator, like `:`
+ ///
+ /// Not implemented in DataFusion yet.
+ Colon,
}
impl Operator {
@@ -188,7 +192,8 @@ impl Operator {
| Operator::AtQuestion
| Operator::Question
| Operator::QuestionAnd
- | Operator::QuestionPipe => None,
+ | Operator::QuestionPipe
+ | Operator::Colon => None,
}
}
@@ -283,7 +288,8 @@ impl Operator {
| Operator::AtQuestion
| Operator::Question
| Operator::QuestionAnd
- | Operator::QuestionPipe => None,
+ | Operator::QuestionPipe
+ | Operator::Colon => None,
}
}
@@ -323,7 +329,8 @@ impl Operator {
| Operator::AtQuestion
| Operator::Question
| Operator::QuestionAnd
- | Operator::QuestionPipe => 30,
+ | Operator::QuestionPipe
+ | Operator::Colon => 30,
Operator::Plus | Operator::Minus => 40,
Operator::Multiply | Operator::Divide | Operator::Modulo => 45,
}
@@ -369,7 +376,8 @@ impl Operator {
| Operator::AtQuestion
| Operator::Question
| Operator::QuestionAnd
- | Operator::QuestionPipe => true,
+ | Operator::QuestionPipe
+ | Operator::Colon => true,
// E.g. `TRUE OR NULL` is `TRUE`
Operator::Or
@@ -429,6 +437,7 @@ impl fmt::Display for Operator {
Operator::Question => "?",
Operator::QuestionAnd => "?&",
Operator::QuestionPipe => "?|",
+ Operator::Colon => ":",
};
write!(f, "{display}")
}
diff --git a/datafusion/expr-common/src/type_coercion/binary.rs
b/datafusion/expr-common/src/type_coercion/binary.rs
index c6ac86cd39..e696545ea6 100644
--- a/datafusion/expr-common/src/type_coercion/binary.rs
+++ b/datafusion/expr-common/src/type_coercion/binary.rs
@@ -324,6 +324,9 @@ impl<'a> BinaryTypeCoercer<'a> {
)
}
},
+ Colon => {
+ Ok(Signature { lhs: lhs.clone(), rhs: rhs.clone(), ret:
lhs.clone() })
+ },
IntegerDivide | Arrow | LongArrow | HashArrow | HashLongArrow
| HashMinus | AtQuestion | Question | QuestionAnd | QuestionPipe => {
not_impl_err!("Operator {} is not yet supported", self.op)
diff --git a/datafusion/physical-expr/src/expressions/binary.rs
b/datafusion/physical-expr/src/expressions/binary.rs
index 72eae396e6..02628b405e 100644
--- a/datafusion/physical-expr/src/expressions/binary.rs
+++ b/datafusion/physical-expr/src/expressions/binary.rs
@@ -715,7 +715,7 @@ impl BinaryExpr {
StringConcat => concat_elements(&left, &right),
AtArrow | ArrowAt | Arrow | LongArrow | HashArrow | HashLongArrow
| AtAt
| HashMinus | AtQuestion | Question | QuestionAnd | QuestionPipe
- | IntegerDivide => {
+ | IntegerDivide | Colon => {
not_impl_err!(
"Binary operator '{:?}' is not supported in the physical
expr",
self.op
diff --git a/datafusion/sql/src/expr/binary_op.rs
b/datafusion/sql/src/expr/binary_op.rs
index edad5bbc6d..4e9025e02e 100644
--- a/datafusion/sql/src/expr/binary_op.rs
+++ b/datafusion/sql/src/expr/binary_op.rs
@@ -22,7 +22,7 @@ use sqlparser::ast::BinaryOperator;
impl<S: ContextProvider> SqlToRel<'_, S> {
pub(crate) fn parse_sql_binary_op(&self, op: &BinaryOperator) ->
Result<Operator> {
- match *op {
+ match op {
BinaryOperator::Gt => Ok(Operator::Gt),
BinaryOperator::GtEq => Ok(Operator::GtEq),
BinaryOperator::Lt => Ok(Operator::Lt),
@@ -68,6 +68,7 @@ impl<S: ContextProvider> SqlToRel<'_, S> {
BinaryOperator::Question => Ok(Operator::Question),
BinaryOperator::QuestionAnd => Ok(Operator::QuestionAnd),
BinaryOperator::QuestionPipe => Ok(Operator::QuestionPipe),
+ BinaryOperator::Custom(s) if s == ":" => Ok(Operator::Colon),
_ => not_impl_err!("Unsupported binary operator: {:?}", op),
}
}
diff --git a/datafusion/sql/src/expr/mod.rs b/datafusion/sql/src/expr/mod.rs
index 9aa5be8131..7902eed1e6 100644
--- a/datafusion/sql/src/expr/mod.rs
+++ b/datafusion/sql/src/expr/mod.rs
@@ -22,8 +22,8 @@ use datafusion_expr::planner::{
use sqlparser::ast::{
AccessExpr, BinaryOperator, CastFormat, CastKind, CeilFloorKind,
DataType as SQLDataType, DateTimeField, DictionaryField, Expr as SQLExpr,
- ExprWithAlias as SQLExprWithAlias, MapEntry, StructField, Subscript,
TrimWhereField,
- TypedString, Value, ValueWithSpan,
+ ExprWithAlias as SQLExprWithAlias, JsonPath, MapEntry, StructField,
Subscript,
+ TrimWhereField, TypedString, Value, ValueWithSpan,
};
use datafusion_common::{
@@ -651,10 +651,36 @@ impl<S: ContextProvider> SqlToRel<'_, S> {
options: Box::new(WildcardOptions::default()),
}),
SQLExpr::Tuple(values) => self.parse_tuple(schema,
planner_context, values),
+ SQLExpr::JsonAccess { value, path } => {
+ self.parse_json_access(schema, planner_context, value, &path)
+ }
_ => not_impl_err!("Unsupported ast node in sqltorel: {sql:?}"),
}
}
+ fn parse_json_access(
+ &self,
+ schema: &DFSchema,
+ planner_context: &mut PlannerContext,
+ value: Box<SQLExpr>,
+ path: &JsonPath,
+ ) -> Result<Expr> {
+ let json_path = path.to_string();
+ let json_path = if let Some(json_path) = json_path.strip_prefix(":") {
+ // sqlparser's JsonPath display adds an extra `:` at the beginning.
+ json_path.to_owned()
+ } else {
+ json_path
+ };
+ self.build_logical_expr(
+ BinaryOperator::Custom(":".to_owned()),
+ self.sql_to_expr(*value, schema, planner_context)?,
+ // pass json path as a string literal, let the impl parse it when
needed.
+ Expr::Literal(ScalarValue::Utf8(Some(json_path)), None),
+ schema,
+ )
+ }
+
/// Parses a struct(..) expression and plans it creation
fn parse_struct(
&self,
diff --git a/datafusion/sql/src/unparser/expr.rs
b/datafusion/sql/src/unparser/expr.rs
index 59a9207b51..b82ab24ade 100644
--- a/datafusion/sql/src/unparser/expr.rs
+++ b/datafusion/sql/src/unparser/expr.rs
@@ -1094,6 +1094,7 @@ impl Unparser<'_> {
Operator::Question => Ok(BinaryOperator::Question),
Operator::QuestionAnd => Ok(BinaryOperator::QuestionAnd),
Operator::QuestionPipe => Ok(BinaryOperator::QuestionPipe),
+ Operator::Colon => Ok(BinaryOperator::Custom(":".to_owned())),
}
}
diff --git a/datafusion/sql/tests/cases/plan_to_sql.rs
b/datafusion/sql/tests/cases/plan_to_sql.rs
index 4717b843ab..670046f164 100644
--- a/datafusion/sql/tests/cases/plan_to_sql.rs
+++ b/datafusion/sql/tests/cases/plan_to_sql.rs
@@ -2821,3 +2821,39 @@ fn test_struct_expr3() {
@r#"SELECT test.c1."metadata".product."name" FROM (SELECT {"metadata":
{product: {"name": 'Product Name'}}} AS c1) AS test"#
);
}
+
+#[test]
+fn test_json_access_1() {
+ let statement = generate_round_trip_statement(
+ GenericDialect {},
+ r#"SELECT j1_string:field FROM j1"#,
+ );
+ assert_snapshot!(
+ statement,
+ @r#"SELECT (j1.j1_string : 'field') FROM j1"#
+ );
+}
+
+#[test]
+fn test_json_access_2() {
+ let statement = generate_round_trip_statement(
+ GenericDialect {},
+ r#"SELECT j1_string:field[0] FROM j1"#,
+ );
+ assert_snapshot!(
+ statement,
+ @r#"SELECT (j1.j1_string : 'field[0]') FROM j1"#
+ );
+}
+
+#[test]
+fn test_json_access_3() {
+ let statement = generate_round_trip_statement(
+ GenericDialect {},
+ r#"SELECT j1_string:field.inner1['inner2'] FROM j1"#,
+ );
+ assert_snapshot!(
+ statement,
+ @r#"SELECT (j1.j1_string : 'field.inner1[''inner2'']') FROM j1"#
+ );
+}
diff --git
a/datafusion/substrait/src/logical_plan/producer/expr/scalar_function.rs
b/datafusion/substrait/src/logical_plan/producer/expr/scalar_function.rs
index bd8a9d9a99..9f70e903a0 100644
--- a/datafusion/substrait/src/logical_plan/producer/expr/scalar_function.rs
+++ b/datafusion/substrait/src/logical_plan/producer/expr/scalar_function.rs
@@ -344,5 +344,6 @@ pub fn operator_to_name(op: Operator) -> &'static str {
Operator::BitwiseXor => "bitwise_xor",
Operator::BitwiseShiftRight => "bitwise_shift_right",
Operator::BitwiseShiftLeft => "bitwise_shift_left",
+ Operator::Colon => "colon",
}
}
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
