Skip to content

fix(arrow/extensions): reject Null Variant typed_value #1243

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
zeroshade merged 3 commits into from
Sep 2, 2026
Merged

fix(arrow/extensions): reject Null Variant typed_value #1243

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
83aed27
fix(arrow/extensions): reject Null Variant typed_value
dgvj-work Aug 28, 2026
d07d9d5
fix(arrow/extensions): handle Null shredded construction
dgvj-work Sep 1, 2026
a5adf0e
fix(arrow/extensions): unwrap nested Variant typed_value extensions
dgvj-work Sep 2, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
43 changes: 35 additions & 8 deletions arrow/extensions/variant.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -122,10 +122,10 @@ func createShreddedField(dt arrow.DataType) arrow.DataType {
// ), Nullable: true})
//
// This is intended to be a convenient way to create a shredded variant type from a definition
// of the fields to shred. If the provided data type is nil, it will create a default
// variant type.
// of the fields to shred. If the provided data type is nil or Null (including an
// extension whose storage is Null), it will create a default variant type.
func NewShreddedVariantType(dt arrow.DataType) *VariantType {
if dt == nil {
if dt == nil || isNullType(dt) {
return NewDefaultVariantType()
}

Expand Down Expand Up @@ -220,9 +220,12 @@ func NewVariantType(storage arrow.DataType) (*VariantType, error) {
return nil, fmt.Errorf("%w: typed_value field must be nullable, got %s", arrow.ErrInvalid, typedValueField.Type)
}

dt := typedValueField.Type
if dt.ID() == arrow.EXTENSION {
dt = dt.(arrow.ExtensionType).StorageType()
dt := storageType(typedValueField.Type)
if dt == nil {
return nil, fmt.Errorf("%w: typed_value field has invalid storage type", arrow.ErrInvalid)
}
if dt.ID() == arrow.NULL {
Comment thread
dgvj-work marked this conversation as resolved.
return nil, fmt.Errorf("%w: typed_value field must not be null type", arrow.ErrInvalid)
}

if nt, ok := dt.(arrow.NestedType); ok {
Expand Down Expand Up @@ -291,11 +294,35 @@ func isBinary(dt arrow.DataType) bool {
dt.ID() == arrow.BINARY_VIEW
}

func storageType(dt arrow.DataType) arrow.DataType {
seen := make(map[arrow.DataType]struct{})
for dt != nil {
ext, ok := dt.(arrow.ExtensionType)
if !ok {
return dt
}
if _, dup := seen[dt]; dup {
return nil
}
seen[dt] = struct{}{}
dt = ext.StorageType()
}
return nil
}

func isNullType(dt arrow.DataType) bool {
st := storageType(dt)
return st == nil || st.ID() == arrow.NULL
}

func validStruct(s *arrow.StructType) bool {
switch s.NumFields() {
case 1:
f := s.Field(0)
return (f.Name == "value" && isBinary(f.Type)) || f.Name == "typed_value"
if f.Name == "value" {
return isBinary(f.Type)
}
return f.Name == "typed_value" && !isNullType(f.Type)
case 2:
valField, ok := s.FieldByName("value")
if !ok || !valField.Nullable || !isBinary(valField.Type) {
Expand All @@ -311,7 +338,7 @@ func validStruct(s *arrow.StructType) bool {
return validNestedType(nt)
}

return true
return !isNullType(typedField.Type)
default:
return false
}
Expand Down
55 changes: 55 additions & 0 deletions arrow/extensions/variant_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -83,6 +83,27 @@ func TestVariantExtensionType(t *testing.T) {
arrow.Field{Name: "metadata", Type: arrow.BinaryTypes.String, Nullable: false},
arrow.Field{Name: "value", Type: arrow.BinaryTypes.Binary, Nullable: false}),
"metadata field must be non-nullable binary type, got utf8"},
{arrow.StructOf(
arrow.Field{Name: "metadata", Type: arrow.BinaryTypes.Binary, Nullable: false},
arrow.Field{Name: "value", Type: arrow.BinaryTypes.Binary, Nullable: true},
arrow.Field{Name: "typed_value", Type: arrow.Null, Nullable: true}),
"typed_value field must not be null type"},
{arrow.StructOf(
arrow.Field{Name: "metadata", Type: arrow.BinaryTypes.Binary, Nullable: false},
arrow.Field{Name: "value", Type: arrow.BinaryTypes.Binary, Nullable: true},
arrow.Field{Name: "typed_value", Type: extensions.NewOpaqueType(arrow.Null, "null", "test"), Nullable: true}),
"typed_value field must not be null type"},
{arrow.StructOf(
arrow.Field{Name: "metadata", Type: arrow.BinaryTypes.Binary, Nullable: false},
arrow.Field{Name: "value", Type: arrow.BinaryTypes.Binary, Nullable: true},
arrow.Field{Name: "typed_value", Type: extensions.NewOpaqueType(
extensions.NewOpaqueType(arrow.Null, "null", "test"), "null", "test"), Nullable: true}),
"typed_value field must not be null type"},
{arrow.StructOf(
arrow.Field{Name: "metadata", Type: arrow.BinaryTypes.Binary, Nullable: false},
arrow.Field{Name: "value", Type: arrow.BinaryTypes.Binary, Nullable: true},
arrow.Field{Name: "typed_value", Type: extensions.NewOpaqueType(nil, "null", "test"), Nullable: true}),
"typed_value field has invalid storage type"},
}

for _, tt := range tests {
Expand Down Expand Up @@ -113,6 +134,31 @@ func TestVariantExtensionBadNestedTypes(t *testing.T) {
), Nullable: false})},
{"empty struct elem", arrow.StructOf(
arrow.Field{Name: "foobar", Type: arrow.StructOf(), Nullable: false})},
{"null typed_value in shredded field", arrow.StructOf(
arrow.Field{Name: "foobar", Type: arrow.StructOf(
arrow.Field{Name: "value", Type: arrow.BinaryTypes.Binary, Nullable: true},
arrow.Field{Name: "typed_value", Type: arrow.Null, Nullable: true},
), Nullable: false})},
{"null typed_value extension in one-field shredded field", arrow.StructOf(
arrow.Field{Name: "foobar", Type: arrow.StructOf(
arrow.Field{Name: "typed_value", Type: extensions.NewOpaqueType(arrow.Null, "null", "test"), Nullable: true},
), Nullable: false})},
{"null typed_value extension in two-field shredded field", arrow.StructOf(
arrow.Field{Name: "foobar", Type: arrow.StructOf(
arrow.Field{Name: "value", Type: arrow.BinaryTypes.Binary, Nullable: true},
arrow.Field{Name: "typed_value", Type: extensions.NewOpaqueType(arrow.Null, "null", "test"), Nullable: true},
), Nullable: false})},
{"double-wrapped null typed_value in one-field shredded field", arrow.StructOf(
arrow.Field{Name: "foobar", Type: arrow.StructOf(
arrow.Field{Name: "typed_value", Type: extensions.NewOpaqueType(
extensions.NewOpaqueType(arrow.Null, "null", "test"), "null", "test"), Nullable: true},
), Nullable: false})},
{"double-wrapped null typed_value in two-field shredded field", arrow.StructOf(
arrow.Field{Name: "foobar", Type: arrow.StructOf(
arrow.Field{Name: "value", Type: arrow.BinaryTypes.Binary, Nullable: true},
arrow.Field{Name: "typed_value", Type: extensions.NewOpaqueType(
extensions.NewOpaqueType(arrow.Null, "null", "test"), "null", "test"), Nullable: true},
), Nullable: false})},
{"non-nullable two elem struct", arrow.StructOf(
arrow.Field{Name: "foobar", Type: arrow.StructOf(
arrow.Field{Name: "value", Type: arrow.BinaryTypes.Binary, Nullable: true},
Expand Down Expand Up @@ -1560,6 +1606,15 @@ func TestVariantBuilderUnmarshalJSON(t *testing.T) {
func TestNewSimpleShreddedVariantType(t *testing.T) {
assert.True(t, arrow.TypeEqual(extensions.NewDefaultVariantType(),
extensions.NewShreddedVariantType(nil)))
assert.True(t, arrow.TypeEqual(extensions.NewDefaultVariantType(),
extensions.NewShreddedVariantType(arrow.Null)))
assert.True(t, arrow.TypeEqual(extensions.NewDefaultVariantType(),
extensions.NewShreddedVariantType(extensions.NewOpaqueType(arrow.Null, "null", "test"))))
assert.True(t, arrow.TypeEqual(extensions.NewDefaultVariantType(),
extensions.NewShreddedVariantType(extensions.NewOpaqueType(
extensions.NewOpaqueType(arrow.Null, "null", "test"), "null", "test"))))
assert.True(t, arrow.TypeEqual(extensions.NewDefaultVariantType(),
extensions.NewShreddedVariantType(extensions.NewOpaqueType(nil, "null", "test"))))

vt := extensions.NewShreddedVariantType(arrow.PrimitiveTypes.Float32)
s := arrow.StructOf(
Expand Down
Loading