vortex_array/arrow/mod.rs
1// SPDX-License-Identifier: Apache-2.0
2// SPDX-FileCopyrightText: Copyright the Vortex contributors
3
4//! Utilities to work with `Arrow` data and types.
5
6use arrow_array::ArrayRef as ArrowArrayRef;
7use arrow_schema::DataType;
8use vortex_error::VortexResult;
9
10mod convert;
11mod datum;
12mod executor;
13mod iter;
14mod null_buffer;
15mod session;
16
17pub(crate) use convert::nulls;
18pub use datum::*;
19pub use executor::*;
20pub use iter::*;
21pub use null_buffer::to_arrow_null_buffer;
22pub use null_buffer::to_null_buffer;
23pub use session::*;
24
25use crate::ArrayRef;
26use crate::LEGACY_SESSION;
27use crate::VortexSessionExecute;
28
29/// Construct a Vortex array from an Arrow array (or other Arrow container) of type `A`.
30///
31/// Implementations reuse the underlying Arrow buffers without copying wherever the Arrow and
32/// Vortex memory layouts allow it.
33pub trait FromArrowArray<A> {
34 /// Convert `array` into a Vortex array whose [`DType`](crate::dtype::DType) has the requested
35 /// `nullable` [`Nullability`](crate::dtype::Nullability).
36 ///
37 /// An Arrow array can carry a validity (null) buffer regardless of whether its schema declares
38 /// the field nullable, so the desired nullability is supplied separately by the caller
39 /// (typically from the corresponding Arrow `Field`'s `is_nullable`). This flag is reconciled
40 /// with the array's physical nulls as follows:
41 ///
42 /// - `nullable == true`: the resulting validity is derived from the array's null buffer, or
43 /// all-valid when the array has none.
44 /// - `nullable == false`: the array must contain no nulls, and the result is non-nullable.
45 ///
46 /// # Errors
47 ///
48 /// Returns an error if `nullable` is `false` but `array` physically contains one or more nulls
49 /// (including an Arrow `NullArray`, which is entirely null), or if the Arrow data type is not
50 /// supported.
51 fn from_arrow(array: A, nullable: bool) -> VortexResult<Self>
52 where
53 Self: Sized;
54}
55
56#[deprecated(note = "Use `execute_arrow(None, ctx)` or `execute_arrow(Some(dt), ctx)` instead")]
57pub trait IntoArrowArray {
58 #[deprecated(note = "Use `execute_arrow(None, ctx)` instead")]
59 fn into_arrow_preferred(self) -> VortexResult<ArrowArrayRef>;
60
61 #[deprecated(note = "Use `execute_arrow(Some(data_type), ctx)` instead")]
62 fn into_arrow(self, data_type: &DataType) -> VortexResult<ArrowArrayRef>;
63}
64
65#[allow(deprecated)]
66impl IntoArrowArray for ArrayRef {
67 /// Convert this [`crate::ArrayRef`] into an Arrow [`crate::ArrayRef`] by using the array's
68 /// preferred (cheapest) Arrow [`DataType`].
69 fn into_arrow_preferred(self) -> VortexResult<ArrowArrayRef> {
70 self.execute_arrow(None, &mut LEGACY_SESSION.create_execution_ctx())
71 }
72
73 fn into_arrow(self, data_type: &DataType) -> VortexResult<ArrowArrayRef> {
74 self.execute_arrow(Some(data_type), &mut LEGACY_SESSION.create_execution_ctx())
75 }
76}