Add a UUID extension type

vortex-array/Cargo.toml

@@ -67,6 +67,7 @@ tabled = { workspace = true, optional = true, default-features = false, features
 ] }
 termtree = { workspace = true }
 tracing = { workspace = true }
+uuid = { workspace = true }
 vortex-buffer = { workspace = true, features = ["arrow"] }
 vortex-error = { workspace = true, features = ["flatbuffers"] }
 vortex-flatbuffers = { workspace = true, features = ["array", "dtype"] }

vortex-array/public-api.lock

@@ -7038,6 +7038,24 @@ pub fn vortex_array::extension::datetime::Timestamp::validate_dtype(&self, _meta
 
 pub fn vortex_array::extension::datetime::Timestamp::validate_scalar_value(&self, metadata: &Self::Metadata, storage_dtype: &vortex_array::dtype::DType, storage_value: &vortex_array::scalar::ScalarValue) -> vortex_error::VortexResult<()>
 
+impl vortex_array::dtype::extension::ExtVTable for vortex_array::extension::uuid::Uuid
+
+pub type vortex_array::extension::uuid::Uuid::Metadata = vortex_array::extension::uuid::UuidMetadata
+
+pub type vortex_array::extension::uuid::Uuid::NativeValue<'a> = uuid::Uuid
+
+pub fn vortex_array::extension::uuid::Uuid::deserialize_metadata(&self, metadata: &[u8]) -> vortex_error::VortexResult<Self::Metadata>
+
+pub fn vortex_array::extension::uuid::Uuid::id(&self) -> vortex_array::dtype::extension::ExtId
+
+pub fn vortex_array::extension::uuid::Uuid::serialize_metadata(&self, metadata: &Self::Metadata) -> vortex_error::VortexResult<alloc::vec::Vec<u8>>
+
+pub fn vortex_array::extension::uuid::Uuid::unpack_native<'a>(&self, metadata: &'a Self::Metadata, _storage_dtype: &'a vortex_array::dtype::DType, storage_value: &'a vortex_array::scalar::ScalarValue) -> vortex_error::VortexResult<Self::NativeValue>
+
+pub fn vortex_array::extension::uuid::Uuid::validate_dtype(&self, _metadata: &Self::Metadata, storage_dtype: &vortex_array::dtype::DType) -> vortex_error::VortexResult<()>
+
+pub fn vortex_array::extension::uuid::Uuid::validate_scalar_value(&self, metadata: &Self::Metadata, storage_dtype: &vortex_array::dtype::DType, storage_value: &vortex_array::scalar::ScalarValue) -> vortex_error::VortexResult<()>
+
 pub trait vortex_array::dtype::extension::Matcher
 
 pub type vortex_array::dtype::extension::Matcher::Match<'a>
@@ -11002,6 +11020,82 @@ pub fn vortex_array::extension::datetime::TimestampOptions::hash<__H: core::hash
 
 impl core::marker::StructuralPartialEq for vortex_array::extension::datetime::TimestampOptions
 
+pub mod vortex_array::extension::uuid
+
+pub struct vortex_array::extension::uuid::Uuid
+
+impl core::clone::Clone for vortex_array::extension::uuid::Uuid
+
+pub fn vortex_array::extension::uuid::Uuid::clone(&self) -> vortex_array::extension::uuid::Uuid
+
+impl core::cmp::Eq for vortex_array::extension::uuid::Uuid
+
+impl core::cmp::PartialEq for vortex_array::extension::uuid::Uuid
+
+pub fn vortex_array::extension::uuid::Uuid::eq(&self, other: &vortex_array::extension::uuid::Uuid) -> bool
+
+impl core::default::Default for vortex_array::extension::uuid::Uuid
+
+pub fn vortex_array::extension::uuid::Uuid::default() -> vortex_array::extension::uuid::Uuid
+
+impl core::fmt::Debug for vortex_array::extension::uuid::Uuid
+
+pub fn vortex_array::extension::uuid::Uuid::fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result
+
+impl core::hash::Hash for vortex_array::extension::uuid::Uuid
+
+pub fn vortex_array::extension::uuid::Uuid::hash<__H: core::hash::Hasher>(&self, state: &mut __H)
+
+impl core::marker::StructuralPartialEq for vortex_array::extension::uuid::Uuid
+
+impl vortex_array::dtype::extension::ExtVTable for vortex_array::extension::uuid::Uuid
+
+pub type vortex_array::extension::uuid::Uuid::Metadata = vortex_array::extension::uuid::UuidMetadata
+
+pub type vortex_array::extension::uuid::Uuid::NativeValue<'a> = uuid::Uuid
+
+pub fn vortex_array::extension::uuid::Uuid::deserialize_metadata(&self, metadata: &[u8]) -> vortex_error::VortexResult<Self::Metadata>
+
+pub fn vortex_array::extension::uuid::Uuid::id(&self) -> vortex_array::dtype::extension::ExtId
+
+pub fn vortex_array::extension::uuid::Uuid::serialize_metadata(&self, metadata: &Self::Metadata) -> vortex_error::VortexResult<alloc::vec::Vec<u8>>
+
+pub fn vortex_array::extension::uuid::Uuid::unpack_native<'a>(&self, metadata: &'a Self::Metadata, _storage_dtype: &'a vortex_array::dtype::DType, storage_value: &'a vortex_array::scalar::ScalarValue) -> vortex_error::VortexResult<Self::NativeValue>
+
+pub fn vortex_array::extension::uuid::Uuid::validate_dtype(&self, _metadata: &Self::Metadata, storage_dtype: &vortex_array::dtype::DType) -> vortex_error::VortexResult<()>
+
+pub fn vortex_array::extension::uuid::Uuid::validate_scalar_value(&self, metadata: &Self::Metadata, storage_dtype: &vortex_array::dtype::DType, storage_value: &vortex_array::scalar::ScalarValue) -> vortex_error::VortexResult<()>
+
+pub struct vortex_array::extension::uuid::UuidMetadata
+
+pub vortex_array::extension::uuid::UuidMetadata::version: core::option::Option<uuid::Version>
+
+impl core::clone::Clone for vortex_array::extension::uuid::UuidMetadata
+
+pub fn vortex_array::extension::uuid::UuidMetadata::clone(&self) -> vortex_array::extension::uuid::UuidMetadata
+
+impl core::cmp::Eq for vortex_array::extension::uuid::UuidMetadata
+
+impl core::cmp::PartialEq for vortex_array::extension::uuid::UuidMetadata
+
+pub fn vortex_array::extension::uuid::UuidMetadata::eq(&self, other: &Self) -> bool
+
+impl core::default::Default for vortex_array::extension::uuid::UuidMetadata
+
+pub fn vortex_array::extension::uuid::UuidMetadata::default() -> vortex_array::extension::uuid::UuidMetadata
+
+impl core::fmt::Debug for vortex_array::extension::uuid::UuidMetadata
+
+pub fn vortex_array::extension::uuid::UuidMetadata::fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result
+
+impl core::fmt::Display for vortex_array::extension::uuid::UuidMetadata
+
+pub fn vortex_array::extension::uuid::UuidMetadata::fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result
+
+impl core::hash::Hash for vortex_array::extension::uuid::UuidMetadata
+
+pub fn vortex_array::extension::uuid::UuidMetadata::hash<H: core::hash::Hasher>(&self, state: &mut H)
+
 pub struct vortex_array::extension::EmptyMetadata
 
 impl core::clone::Clone for vortex_array::extension::EmptyMetadata

vortex-array/src/extension/mod.rs

@@ -6,6 +6,7 @@
 use std::fmt;
 
 pub mod datetime;
+pub mod uuid;
 
 #[cfg(test)]
 mod tests;

vortex-array/src/extension/uuid/metadata.rs

@@ -0,0 +1,63 @@
+// SPDX-License-Identifier: Apache-2.0
+// SPDX-FileCopyrightText: Copyright the Vortex contributors
+
+use std::fmt;
+use std::hash::Hash;
+use std::hash::Hasher;
+
+use uuid::Version;
+use vortex_error::VortexResult;
+use vortex_error::vortex_bail;
+
+/// Converts a `u8` discriminant back to a [`uuid::Version`].
+pub(crate) fn u8_to_version(b: u8) -> VortexResult<Version> {
+    match b {
+        0 => Ok(Version::Nil),
+        1 => Ok(Version::Mac),
+        2 => Ok(Version::Dce),
+        3 => Ok(Version::Md5),
+        4 => Ok(Version::Random),
+        5 => Ok(Version::Sha1),
+        6 => Ok(Version::SortMac),
+        7 => Ok(Version::SortRand),
+        8 => Ok(Version::Custom),
+        0xff => Ok(Version::Max),
+        _ => vortex_bail!("unknown UUID version discriminant: {b}"),
+    }
+}
+
+/// Metadata for the UUID extension type.
+///
+/// Optionally records which UUID version the column contains (e.g. v4 random, v7
+/// sort-random). When `None`, the column may contain any mix of versions.
+#[derive(Clone, Debug, Default)]
+pub struct UuidMetadata {
+    /// The UUID version, if known.
+    pub version: Option<Version>,
+}
+
+impl fmt::Display for UuidMetadata {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        match self.version {
+            None => write!(f, ""),
+            Some(v) => write!(f, "v{}", v as u8),
+        }
+    }
+}
+
+// `uuid::Version` derives `PartialEq` but not `Eq` or `Hash`, so we implement these
+// manually using the `#[repr(u8)]` discriminant.
+
+impl PartialEq for UuidMetadata {
+    fn eq(&self, other: &Self) -> bool {
+        self.version.map(|v| v as u8) == other.version.map(|v| v as u8)
+    }
+}
+
+impl Eq for UuidMetadata {}
+
+impl Hash for UuidMetadata {
+    fn hash<H: Hasher>(&self, state: &mut H) {
+        self.version.map(|v| v as u8).hash(state);
+    }
+}

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

@@ -0,0 +1,20 @@
+// SPDX-License-Identifier: Apache-2.0
+// SPDX-FileCopyrightText: Copyright the Vortex contributors
+
+//! UUID extension type for Vortex.
+//!
+//! Provides a UUID extension type backed by `FixedSizeList(Primitive(U8), 16)` storage. Each UUID
+//! is stored as 16 bytes in big-endian (network) byte order, matching [RFC 4122] and Arrow's
+//! [canonical UUID extension].
+//!
+//! [RFC 4122]: https://www.rfc-editor.org/rfc/rfc4122
+//! [canonical UUID extension]: https://arrow.apache.org/docs/format/CanonicalExtensions.html#uuid
+
+mod metadata;
+pub use metadata::UuidMetadata;
+
+pub(crate) mod vtable;
+
+/// The VTable for the UUID extension type.
+#[derive(Clone, Debug, Default, PartialEq, Eq, Hash)]
+pub struct Uuid;