Extension Scalar VTable

vortex-array/src/arrays/extension/compute/rules.rs

@@ -75,23 +75,42 @@ mod tests {
     use crate::extension::EmptyMetadata;
     use crate::optimizer::ArrayOptimizer;
     use crate::scalar::Scalar;
+    use crate::scalar::ScalarValue;
 
     #[derive(Clone, Debug, Default, PartialEq, Eq, Hash)]
     struct TestExt;
     impl ExtVTable for TestExt {
         type Metadata = EmptyMetadata;
+        type NativeValue<'a> = &'a str;
 
         fn id(&self) -> ExtId {
             ExtId::new_ref("test_ext")
         }
 
+        fn serialize_metadata(&self, _metadata: &Self::Metadata) -> VortexResult<Vec<u8>> {
+            Ok(vec![])
+        }
+
+        fn deserialize_metadata(&self, _data: &[u8]) -> VortexResult<Self::Metadata> {
+            Ok(EmptyMetadata)
+        }
+
         fn validate_dtype(
             &self,
             _options: &Self::Metadata,
             _storage_dtype: &DType,
         ) -> VortexResult<()> {
             Ok(())
         }
+
+        fn unpack_native<'a>(
+            &self,
+            _metadata: &'a Self::Metadata,
+            _storage_dtype: &'a DType,
+            _storage_value: &'a ScalarValue,
+        ) -> VortexResult<Self::NativeValue<'a>> {
+            Ok("")
+        }
     }
 
     fn test_ext_dtype() -> ExtDTypeRef {
@@ -164,18 +183,36 @@ mod tests {
         struct TestExt2;
         impl ExtVTable for TestExt2 {
             type Metadata = EmptyMetadata;
+            type NativeValue<'a> = &'a str;
 
             fn id(&self) -> ExtId {
                 ExtId::new_ref("test_ext_2")
             }
 
+            fn serialize_metadata(&self, _metadata: &Self::Metadata) -> VortexResult<Vec<u8>> {
+                Ok(vec![])
+            }
+
+            fn deserialize_metadata(&self, _data: &[u8]) -> VortexResult<Self::Metadata> {
+                Ok(EmptyMetadata)
+            }
+
             fn validate_dtype(
                 &self,
                 _options: &Self::Metadata,
                 _storage_dtype: &DType,
             ) -> VortexResult<()> {
                 Ok(())
             }
+
+            fn unpack_native<'a>(
+                &self,
+                _metadata: &'a Self::Metadata,
+                _storage_dtype: &'a DType,
+                _storage_value: &'a ScalarValue,
+            ) -> VortexResult<Self::NativeValue<'a>> {
+                Ok("")
+            }
         }
 
         let ext_dtype1 = ExtDType::<TestExt>::try_new(

vortex-array/src/dtype/extension/mod.rs

@@ -1,7 +1,7 @@
 // SPDX-License-Identifier: Apache-2.0
 // SPDX-FileCopyrightText: Copyright the Vortex contributors
 
-//! Extension DTypes, and interfaces for working with extension types (dtypes, scalars, and arrays).
+//! Extension DTypes, and interfaces for working with extension types.
 //!
 //! ## File layout convention
 //!

vortex-array/src/dtype/extension/plugin.rs

@@ -32,7 +32,7 @@ impl<V: ExtVTable> ExtDTypePlugin for V {
     }
 
     fn deserialize(&self, data: &[u8], storage_dtype: DType) -> VortexResult<ExtDTypeRef> {
-        let metadata = ExtVTable::deserialize(self, data)?;
+        let metadata = ExtVTable::deserialize_metadata(self, data)?;
         Ok(ExtDType::try_with_vtable(self.clone(), metadata, storage_dtype)?.erased())
     }
 }

vortex-array/src/dtype/extension/typed.rs

@@ -168,7 +168,7 @@ impl<V: ExtVTable> DynExtDType for ExtDTypeInner<V> {
     }
 
     fn metadata_serialize(&self) -> VortexResult<Vec<u8>> {
-        V::serialize(&self.vtable, &self.metadata)
+        V::serialize_metadata(&self.vtable, &self.metadata)
     }
 
     fn with_nullability(&self, nullability: Nullability) -> ExtDTypeRef {

vortex-array/src/dtype/extension/vtable.rs

@@ -6,40 +6,70 @@ use std::fmt::Display;
 use std::hash::Hash;
 
 use vortex_error::VortexResult;
-use vortex_error::vortex_bail;
 
 use crate::dtype::DType;
 use crate::dtype::extension::ExtId;
+use crate::scalar::ScalarValue;
 
 /// The public API for defining new extension types.
 ///
 /// This is the non-object-safe trait that plugin authors implement to define a new extension
 /// type. It specifies the type's identity, metadata, serialization, and validation.
 pub trait ExtVTable: 'static + Sized + Send + Sync + Clone + Debug + Eq + Hash {
-    /// Associated type containing the deserialized metadata for this extension type
+    /// Associated type containing the deserialized metadata for this extension type.
     type Metadata: 'static + Send + Sync + Clone + Debug + Display + Eq + Hash;
 
+    /// A native Rust value that represents a scalar of the extension type.
+    ///
+    /// The value only represents non-null values. We denote nullable values as `Option<Value>`.
+    type NativeValue<'a>: Display;
+
     /// Returns the ID for this extension type.
     fn id(&self) -> ExtId;
 
+    // Methods related to the extension `DType`.
+
     /// Serialize the metadata into a byte vector.
-    fn serialize(&self, metadata: &Self::Metadata) -> VortexResult<Vec<u8>> {
-        _ = metadata;
-        vortex_bail!(
-            "Serialization not implemented for extension type {}",
-            self.id()
-        );
-    }
+    fn serialize_metadata(&self, metadata: &Self::Metadata) -> VortexResult<Vec<u8>>;
 
     /// Deserialize the metadata from a byte slice.
-    fn deserialize(&self, metadata: &[u8]) -> VortexResult<Self::Metadata> {
-        _ = metadata;
-        vortex_bail!(
-            "Deserialization not implemented for extension type {}",
-            self.id()
-        );
-    }
+    fn deserialize_metadata(&self, metadata: &[u8]) -> VortexResult<Self::Metadata>;
 
     /// Validate that the given storage type is compatible with this extension type.
     fn validate_dtype(&self, metadata: &Self::Metadata, storage_dtype: &DType) -> VortexResult<()>;
+
+    // Methods related to the extension scalar values.
+
+    /// Validate the given storage value is compatible with the extension type.
+    ///
+    /// By default, this calls [`unpack_native()`](ExtVTable::unpack_native) and discards the result.
+    ///
+    /// # Errors
+    ///
+    /// Returns an error if the storage [`ScalarValue`] is not compatible with the extension type.
+    fn validate_scalar_value(
+        &self,
+        metadata: &Self::Metadata,
+        storage_dtype: &DType,
+        storage_value: &ScalarValue,
+    ) -> VortexResult<()> {
+        self.unpack_native(metadata, storage_dtype, storage_value)
+            .map(|_| ())
+    }
+
+    /// Validate and unpack a native value from the storage [`ScalarValue`].
+    ///
+    /// Note that [`ExtVTable::validate_dtype()`] is always called first to validate the storage
+    /// [`DType`], and the [`Scalar`](crate::scalar::Scalar) implementation will verify that the
+    /// storage value is compatible with the storage dtype on construction.
+    ///
+    /// # Errors
+    ///
+    /// Returns an error if the storage [`ScalarValue`] is not compatible with the extension type.
+    fn unpack_native<'a>(
+        &self,
+        metadata: &'a Self::Metadata,
+        storage_dtype: &'a DType,
+        storage_value: &'a ScalarValue,
+    ) -> VortexResult<Self::NativeValue<'a>>;
 }

vortex-array/src/extension/datetime/date.rs

@@ -1,8 +1,12 @@
 // SPDX-License-Identifier: Apache-2.0
 // SPDX-FileCopyrightText: Copyright the Vortex contributors
 
+use std::fmt;
+
+use jiff::Span;
 use vortex_error::VortexExpect;
 use vortex_error::VortexResult;
+use vortex_error::vortex_bail;
 use vortex_error::vortex_ensure;
 use vortex_error::vortex_err;
 
@@ -13,11 +17,25 @@ use crate::dtype::extension::ExtDType;
 use crate::dtype::extension::ExtId;
 use crate::dtype::extension::ExtVTable;
 use crate::extension::datetime::TimeUnit;
+use crate::scalar::ScalarValue;
+
+/// The Unix epoch date (1970-01-01).
+const EPOCH: jiff::civil::Date = jiff::civil::Date::constant(1970, 1, 1);
 
 /// Date DType.
 #[derive(Clone, Debug, Default, PartialEq, Eq, Hash)]
 pub struct Date;
 
+fn date_ptype(time_unit: &TimeUnit) -> Option<PType> {
+    match time_unit {
+        TimeUnit::Nanoseconds => None,
+        TimeUnit::Microseconds => None,
+        TimeUnit::Milliseconds => Some(PType::I64),
+        TimeUnit::Seconds => None,
+        TimeUnit::Days => Some(PType::I32),
+    }
+}
+
 impl Date {
     /// Creates a new Date extension dtype with the given time unit and nullability.
     ///
@@ -38,18 +56,37 @@ impl Date {
     }
 }
 
+/// Unpacked value of a [`Date`] extension scalar.
+pub enum DateValue {
+    /// Days since the Unix epoch.
+    Days(i32),
+    /// Milliseconds since the Unix epoch.
+    Milliseconds(i64),
+}
+
+impl fmt::Display for DateValue {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        let date = match self {
+            DateValue::Days(days) => EPOCH + Span::new().days(*days),
+            DateValue::Milliseconds(ms) => EPOCH + Span::new().milliseconds(*ms),
+        };
+        write!(f, "{}", date)
+    }
+}
+
 impl ExtVTable for Date {
     type Metadata = TimeUnit;
+    type NativeValue<'a> = DateValue;
 
     fn id(&self) -> ExtId {
         ExtId::new_ref("vortex.date")
     }
 
-    fn serialize(&self, metadata: &Self::Metadata) -> VortexResult<Vec<u8>> {
+    fn serialize_metadata(&self, metadata: &Self::Metadata) -> VortexResult<Vec<u8>> {
         Ok(vec![u8::from(*metadata)])
     }
 
-    fn deserialize(&self, metadata: &[u8]) -> VortexResult<Self::Metadata> {
+    fn deserialize_metadata(&self, metadata: &[u8]) -> VortexResult<Self::Metadata> {
         let tag = metadata[0];
         TimeUnit::try_from(tag)
     }
@@ -67,14 +104,19 @@ impl ExtVTable for Date {
 
         Ok(())
     }
-}
 
-fn date_ptype(time_unit: &TimeUnit) -> Option<PType> {
-    match time_unit {
-        TimeUnit::Nanoseconds => None,
-        TimeUnit::Microseconds => None,
-        TimeUnit::Milliseconds => Some(PType::I64),
-        TimeUnit::Seconds => None,
-        TimeUnit::Days => Some(PType::I32),
+    fn unpack_native(
+        &self,
+        metadata: &Self::Metadata,
+        _storage_dtype: &DType,
+        storage_value: &ScalarValue,
+    ) -> VortexResult<Self::NativeValue<'_>> {
+        match metadata {
+            TimeUnit::Milliseconds => Ok(DateValue::Milliseconds(
+                storage_value.as_primitive().cast::<i64>()?,
+            )),
+            TimeUnit::Days => Ok(DateValue::Days(storage_value.as_primitive().cast::<i32>()?)),
+            _ => vortex_bail!("Date type does not support time unit {}", metadata),
+        }
     }
 }