Add DType::Union variant carrying just Nullability

This is the first of two commits splitting the original Union work. It adds the
`Union(Nullability)` enum variant, plumbs it through all existing match arms
with `todo!("TODO(connor)[Union]: unimplemented")`, and reserves slot 12 in the
flatbuffer and protobuf wire schemas. A follow-up commit will replace
`Union(Nullability)` with `Union(UnionVariants, Nullability)` and fill in the
serde paths.

Signed-off-by: Connor Tsui <connor.tsui20@gmail.com>

Author: connortsui20

encodings/sparse/src/canonical.rs

@@ -122,6 +122,7 @@ pub(super) fn execute_sparse(parts: SparseParts, ctx: &mut ExecutionCtx) -> Vort
             len,
             ctx,
         )?,
+        DType::Union(..) => todo!("TODO(connor)[Union]: unimplemented"),
         DType::Decimal(decimal_dtype, nullability) => {
             let canonical_decimal_value_type =
                 DecimalType::smallest_decimal_value_type(decimal_dtype);

fuzz/src/array/compare.rs

@@ -173,7 +173,7 @@ pub fn compare_canonical_array(
             }))
             .into_array()
         }
-        d @ (DType::Null | DType::Extension(_) | DType::Variant(_)) => {
+        d @ (DType::Null | DType::Union(..) | DType::Extension(_) | DType::Variant(_)) => {
             unreachable!("DType {d} not supported for fuzzing")
         }
     }

fuzz/src/array/filter.rs

@@ -122,7 +122,7 @@ pub fn filter_canonical_array(
             }
             take_canonical_array_non_nullable_indices(array, indices.as_slice(), ctx)
         }
-        d @ (DType::Null | DType::Extension(_) | DType::Variant(_)) => {
+        d @ (DType::Null | DType::Union(..) | DType::Extension(_) | DType::Variant(_)) => {
             unreachable!("DType {d} not supported for fuzzing")
         }
     }

fuzz/src/array/mod.rs

@@ -474,6 +474,7 @@ fn actions_for_dtype(dtype: &DType) -> HashSet<ActionType> {
                     acc.intersection(&actions).copied().collect()
                 })
         }
+        DType::Union(..) => todo!("TODO(connor)[Union]: unimplemented"),
         DType::List(..) | DType::FixedSizeList(..) => {
             // List supports: Compress, Slice, Take, Filter, MinMax, Mask, ScalarAt
             // Does NOT support: SearchSorted, Compare, Cast, Sum, FillNull

fuzz/src/array/search_sorted.rs

@@ -148,7 +148,7 @@ pub fn search_sorted_canonical_array(
                 .collect::<VortexResult<Vec<_>>>()?;
             scalar_vals.search_sorted(&scalar.cast(array.dtype())?, side)
         }
-        d @ (DType::Null | DType::Extension(_) | DType::Variant(_)) => {
+        d @ (DType::Null | DType::Union(..) | DType::Extension(_) | DType::Variant(_)) => {
             unreachable!("DType {d} not supported for fuzzing")
         }
     }

fuzz/src/array/slice.rs

@@ -123,7 +123,7 @@ pub fn slice_canonical_array(
                 .into_array(),
             )
         }
-        d @ (DType::Null | DType::Extension(_) | DType::Variant(_)) => {
+        d @ (DType::Null | DType::Union(..) | DType::Extension(_) | DType::Variant(_)) => {
             unreachable!("DType {d} not supported for fuzzing")
         }
     }

fuzz/src/array/sort.rs

@@ -101,7 +101,7 @@ pub fn sort_canonical_array(array: &ArrayRef, ctx: &mut ExecutionCtx) -> VortexR
             });
             take_canonical_array_non_nullable_indices(array, &sort_indices, ctx)
         }
-        d @ (DType::Null | DType::Extension(_) | DType::Variant(_)) => {
+        d @ (DType::Null | DType::Union(..) | DType::Extension(_) | DType::Variant(_)) => {
             unreachable!("DType {d} not supported for fuzzing")
         }
     }

fuzz/src/array/take.rs

@@ -147,7 +147,7 @@ pub fn take_canonical_array(
             }
             Ok(builder.finish())
         }
-        d @ (DType::Null | DType::Extension(_) | DType::Variant(_)) => {
+        d @ (DType::Null | DType::Union(..) | DType::Extension(_) | DType::Variant(_)) => {
             unreachable!("DType {d} not supported for fuzzing")
         }
     }

vortex-array/public-api.lock

@@ -9416,6 +9416,8 @@ pub vortex_array::dtype::DType::Primitive(vortex_array::dtype::PType, vortex_arr
 
 pub vortex_array::dtype::DType::Struct(vortex_array::dtype::StructFields, vortex_array::dtype::Nullability)
 
+pub vortex_array::dtype::DType::Union(vortex_array::dtype::Nullability)
+
 pub vortex_array::dtype::DType::Utf8(vortex_array::dtype::Nullability)
 
 pub vortex_array::dtype::DType::Variant(vortex_array::dtype::Nullability)
@@ -9492,6 +9494,8 @@ pub fn vortex_array::dtype::DType::is_signed_int(&self) -> bool
 
 pub fn vortex_array::dtype::DType::is_struct(&self) -> bool
 
+pub fn vortex_array::dtype::DType::is_union(&self) -> bool
+
 pub fn vortex_array::dtype::DType::is_unsigned_int(&self) -> bool
 
 pub fn vortex_array::dtype::DType::is_utf8(&self) -> bool

vortex-array/src/aggregate_fn/fns/uncompressed_size_in_bytes/mod.rs

@@ -238,6 +238,7 @@ pub(crate) fn constant_uncompressed_size_in_bytes(
             let canonical = array.array().clone().execute::<Canonical>(ctx)?;
             return canonical_uncompressed_size_in_bytes(&canonical, ctx);
         }
+        DType::Union(..) => todo!("TODO(connor)[Union]: unimplemented"),
     };
 
     value_size
@@ -284,6 +285,7 @@ fn supports_uncompressed_size_in_bytes(dtype: &DType) -> bool {
         DType::Struct(fields, _) => fields
             .fields()
             .all(|field| supports_uncompressed_size_in_bytes(&field)),
+        DType::Union(_) => todo!("TODO(connor)[Union]: unimplemented"),
         DType::Extension(ext_dtype) => {
             supports_uncompressed_size_in_bytes(ext_dtype.storage_dtype())
         }