11 major breaking releases
35.0.0 | Mar 13, 2023 |
---|---|
34.0.0 | Feb 27, 2023 |
33.0.0 | Feb 13, 2023 |
32.0.0 | Jan 30, 2023 |
24.0.0 | Oct 3, 2022 |
#56 in Algorithms
99,394 downloads per month
Used in 80 crates
(14 directly)
1.5MB
29K
SLoC
The central type in Apache Arrow are arrays, which are a known-length sequence of values
all having the same type. This crate provides concrete implementations of each type, as
well as an [Array
] trait that can be used for type-erasure.
Downcasting an Array
Arrays are often passed around as a dynamically typed &dyn Array
or [ArrayRef
].
For example, RecordBatch
stores columns as [ArrayRef
].
Whilst these arrays can be passed directly to the compute
, csv
, json
, etc... APIs,
it is often the case that you wish to interact with the data directly.
This requires downcasting to the concrete type of the array:
# use arrow_array::{Array, Float32Array, Int32Array};
fn sum_int32(array: &dyn Array) -> i32 {
let integers: &Int32Array = array.as_any().downcast_ref().unwrap();
integers.iter().map(|val| val.unwrap_or_default()).sum()
}
// Note: the values for positions corresponding to nulls will be arbitrary
fn as_f32_slice(array: &dyn Array) -> &[f32] {
array.as_any().downcast_ref::<Float32Array>().unwrap().values()
}
Additionally, there are convenient functions to do this casting
such as [cast::as_primitive_array<T>
] and [cast::as_string_array
]:
# use arrow_array::Array;
# use arrow_array::cast::as_primitive_array;
# use arrow_array::types::Float32Type;
fn as_f32_slice(array: &dyn Array) -> &[f32] {
// use as_primtive_array
as_primitive_array::<Float32Type>(array).values()
}
Building an Array
Most [Array
] implementations can be constructed directly from iterators or [Vec
]
# use arrow_array::{Int32Array, ListArray, StringArray};
# use arrow_array::types::Int32Type;
Int32Array::from(vec![1, 2]);
Int32Array::from(vec![Some(1), None]);
Int32Array::from_iter([1, 2, 3, 4]);
Int32Array::from_iter([Some(1), Some(2), None, Some(4)]);
StringArray::from(vec!["foo", "bar"]);
StringArray::from(vec![Some("foo"), None]);
StringArray::from_iter([Some("foo"), None]);
StringArray::from_iter_values(["foo", "bar"]);
ListArray::from_iter_primitive::<Int32Type, _, _>([
Some(vec![Some(1), None, Some(3)]),
None,
Some(vec![])
]);
Additionally ArrayBuilder
implementations can be
used to construct arrays with a push-based interface
# use arrow_array::Int16Array;
#
// Create a new builder with a capacity of 100
let mut builder = Int16Array::builder(100);
// Append a single primitive value
builder.append_value(1);
// Append a null value
builder.append_null();
// Append a slice of primitive values
builder.append_slice(&[2, 3, 4]);
// Build the array
let array = builder.finish();
assert_eq!(
5,
array.len(),
"The array has 5 values, counting the null value"
);
assert_eq!(2, array.value(2), "Get the value with index 2");
assert_eq!(
&array.values()[3..5],
&[3, 4],
"Get slice of len 2 starting at idx 3"
)
Zero-Copy Slicing
Given an [Array
] of arbitrary length, it is possible to create an owned slice of this
data. Internally this just increments some ref-counts, and so is incredibly cheap
# use std::sync::Arc;
# use arrow_array::{ArrayRef, Int32Array};
let array = Arc::new(Int32Array::from_iter([1, 2, 3])) as ArrayRef;
// Slice with offset 1 and length 2
let sliced = array.slice(1, 2);
let ints = sliced.as_any().downcast_ref::<Int32Array>().unwrap();
assert_eq!(ints.values(), &[2, 3]);
Internal Representation
Internally, arrays are represented by one or several Buffer
, the number and meaning of
which depend on the array’s data type, as documented in the Arrow specification.
For example, the type [Int16Array
] represents an array of 16-bit integers and consists of:
- An optional
NullBuffer
identifying any null values - A contiguous
Buffer
of 16-bit integers
Similarly, the type [StringArray
] represents an array of UTF-8 strings and consists of:
- An optional
NullBuffer
identifying any null values - An offsets
Buffer
of 32-bit integers identifying valid UTF-8 sequences within the values buffer - A values
Buffer
of UTF-8 encoded string data
Dependencies
~2–28MB
~469K SLoC