This repository has been archived by the owner on Feb 18, 2024. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 221
/
record_batch.rs
334 lines (313 loc) · 11.3 KB
/
record_batch.rs
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
//! Contains [`RecordBatch`].
use std::sync::Arc;
use crate::array::*;
use crate::datatypes::*;
use crate::error::{ArrowError, Result};
/// A two-dimensional dataset with a number of
/// columns ([`Array`]) and rows and defined [`Schema`](crate::datatypes::Schema).
/// # Implementation
/// Cloning is `O(C)` where `C` is the number of columns.
#[derive(Clone, Debug, PartialEq)]
pub struct RecordBatch {
schema: Arc<Schema>,
columns: Vec<Arc<dyn Array>>,
}
impl RecordBatch {
/// Creates a [`RecordBatch`] from a schema and columns.
/// # Errors
/// This function errors iff
/// * `columns` is empty
/// * the schema and column data types do not match
/// * `columns` have a different length
/// # Example
///
/// ```
/// # use std::sync::Arc;
/// # use arrow2::array::PrimitiveArray;
/// # use arrow2::datatypes::{Schema, Field, DataType};
/// # use arrow2::record_batch::RecordBatch;
/// # fn main() -> arrow2::error::Result<()> {
/// let id_array = PrimitiveArray::from_slice([1i32, 2, 3, 4, 5]);
/// let schema = Arc::new(Schema::new(vec![
/// Field::new("id", DataType::Int32, false)
/// ]));
///
/// let batch = RecordBatch::try_new(
/// schema,
/// vec![Arc::new(id_array)]
/// )?;
/// # Ok(())
/// # }
/// ```
pub fn try_new(schema: Arc<Schema>, columns: Vec<Arc<dyn Array>>) -> Result<Self> {
let options = RecordBatchOptions::default();
Self::validate_new_batch(&schema, columns.as_slice(), &options)?;
Ok(RecordBatch { schema, columns })
}
/// Creates a [`RecordBatch`] from a schema and columns, with additional options,
/// such as whether to strictly validate field names.
///
/// See [`Self::try_new()`] for the expected conditions.
pub fn try_new_with_options(
schema: Arc<Schema>,
columns: Vec<Arc<dyn Array>>,
options: &RecordBatchOptions,
) -> Result<Self> {
Self::validate_new_batch(&schema, &columns, options)?;
Ok(RecordBatch { schema, columns })
}
/// Creates a new empty [`RecordBatch`].
pub fn new_empty(schema: Arc<Schema>) -> Self {
let columns = schema
.fields()
.iter()
.map(|field| new_empty_array(field.data_type().clone()).into())
.collect();
RecordBatch { schema, columns }
}
/// Validate the schema and columns using [`RecordBatchOptions`]. Returns an error
/// if any validation check fails.
fn validate_new_batch(
schema: &Schema,
columns: &[Arc<dyn Array>],
options: &RecordBatchOptions,
) -> Result<()> {
// check that there are some columns
if columns.is_empty() {
return Err(ArrowError::InvalidArgumentError(
"at least one column must be defined to create a record batch".to_string(),
));
}
// check that number of fields in schema match column length
if schema.fields().len() != columns.len() {
return Err(ArrowError::InvalidArgumentError(format!(
"number of columns({}) must match number of fields({}) in schema",
columns.len(),
schema.fields().len(),
)));
}
// check that all columns have the same row count, and match the schema
let len = columns[0].len();
// This is a bit repetitive, but it is better to check the condition outside the loop
if options.match_field_names {
for (i, column) in columns.iter().enumerate() {
if column.len() != len {
return Err(ArrowError::InvalidArgumentError(
"all columns in a record batch must have the same length".to_string(),
));
}
if column.data_type() != schema.field(i).data_type() {
return Err(ArrowError::InvalidArgumentError(format!(
"column types must match schema types, expected {:?} but found {:?} at column index {}",
schema.field(i).data_type(),
column.data_type(),
i)));
}
}
} else {
for (i, column) in columns.iter().enumerate() {
if column.len() != len {
return Err(ArrowError::InvalidArgumentError(
"all columns in a record batch must have the same length".to_string(),
));
}
if !column
.data_type()
.equals_datatype(schema.field(i).data_type())
{
return Err(ArrowError::InvalidArgumentError(format!(
"column types must match schema types, expected {:?} but found {:?} at column index {}",
schema.field(i).data_type(),
column.data_type(),
i)));
}
}
}
Ok(())
}
/// Returns the [`Schema`](crate::datatypes::Schema) of the record batch.
pub fn schema(&self) -> &Arc<Schema> {
&self.schema
}
/// Returns the number of columns in the record batch.
///
/// # Example
///
/// ```
/// # use std::sync::Arc;
/// # use arrow2::array::PrimitiveArray;
/// # use arrow2::datatypes::{Schema, Field, DataType};
/// # use arrow2::record_batch::RecordBatch;
/// # fn main() -> arrow2::error::Result<()> {
/// let id_array = PrimitiveArray::from_slice([1i32, 2, 3, 4, 5]);
/// let schema = Arc::new(Schema::new(vec![
/// Field::new("id", DataType::Int32, false)
/// ]));
///
/// let batch = RecordBatch::try_new(schema, vec![Arc::new(id_array)])?;
///
/// assert_eq!(batch.num_columns(), 1);
/// # Ok(())
/// # }
/// ```
pub fn num_columns(&self) -> usize {
self.columns.len()
}
/// Returns the number of rows in each column.
///
/// # Panics
///
/// Panics if the `RecordBatch` contains no columns.
///
/// # Example
///
/// ```
/// # use std::sync::Arc;
/// # use arrow2::array::PrimitiveArray;
/// # use arrow2::datatypes::{Schema, Field, DataType};
/// # use arrow2::record_batch::RecordBatch;
/// # fn main() -> arrow2::error::Result<()> {
/// let id_array = PrimitiveArray::from_slice([1i32, 2, 3, 4, 5]);
/// let schema = Arc::new(Schema::new(vec![
/// Field::new("id", DataType::Int32, false)
/// ]));
///
/// let batch = RecordBatch::try_new(schema, vec![Arc::new(id_array)])?;
///
/// assert_eq!(batch.num_rows(), 5);
/// # Ok(())
/// # }
/// ```
pub fn num_rows(&self) -> usize {
self.columns[0].len()
}
/// Get a reference to a column's array by index.
///
/// # Panics
///
/// Panics if `index` is outside of `0..num_columns`.
pub fn column(&self, index: usize) -> &Arc<dyn Array> {
&self.columns[index]
}
/// Get a reference to all columns in the record batch.
pub fn columns(&self) -> &[Arc<dyn Array>] {
&self.columns[..]
}
/// Create a `RecordBatch` from an iterable list of pairs of the
/// form `(field_name, array)`, with the same requirements on
/// fields and arrays as [`RecordBatch::try_new`]. This method is
/// often used to create a single `RecordBatch` from arrays,
/// e.g. for testing.
///
/// The resulting schema is marked as nullable for each column if
/// the array for that column is has any nulls. To explicitly
/// specify nullibility, use [`RecordBatch::try_from_iter_with_nullable`]
///
/// Example:
/// ```
/// use std::sync::Arc;
/// use arrow2::array::*;
/// use arrow2::datatypes::DataType;
/// use arrow2::record_batch::RecordBatch;
///
/// let a: Arc<dyn Array> = Arc::new(Int32Array::from_slice(&[1, 2]));
/// let b: Arc<dyn Array> = Arc::new(Utf8Array::<i32>::from_slice(&["a", "b"]));
///
/// let record_batch = RecordBatch::try_from_iter(vec![
/// ("a", a),
/// ("b", b),
/// ]);
/// ```
pub fn try_from_iter<I, F>(value: I) -> Result<Self>
where
I: IntoIterator<Item = (F, Arc<dyn Array>)>,
F: AsRef<str>,
{
// TODO: implement `TryFrom` trait, once
// https://github.com/rust-lang/rust/issues/50133 is no longer an
// issue
let iter = value.into_iter().map(|(field_name, array)| {
let nullable = array.null_count() > 0;
(field_name, array, nullable)
});
Self::try_from_iter_with_nullable(iter)
}
/// Create a `RecordBatch` from an iterable list of tuples of the
/// form `(field_name, array, nullable)`, with the same requirements on
/// fields and arrays as [`RecordBatch::try_new`]. This method is often
/// used to create a single `RecordBatch` from arrays, e.g. for
/// testing.
///
/// Example:
/// ```
/// use std::sync::Arc;
/// use arrow2::array::*;
/// use arrow2::datatypes::DataType;
/// use arrow2::record_batch::RecordBatch;
///
/// let a: Arc<dyn Array> = Arc::new(Int32Array::from_slice(&[1, 2]));
/// let b: Arc<dyn Array> = Arc::new(Utf8Array::<i32>::from_slice(&["a", "b"]));
///
/// // Note neither `a` nor `b` has any actual nulls, but we mark
/// // b an nullable
/// let record_batch = RecordBatch::try_from_iter_with_nullable(vec![
/// ("a", a, false),
/// ("b", b, true),
/// ]);
/// ```
pub fn try_from_iter_with_nullable<I, F>(value: I) -> Result<Self>
where
I: IntoIterator<Item = (F, Arc<dyn Array>, bool)>,
F: AsRef<str>,
{
// TODO: implement `TryFrom` trait, once
// https://github.com/rust-lang/rust/issues/50133 is no longer an
// issue
let (fields, columns) = value
.into_iter()
.map(|(field_name, array, nullable)| {
let field_name = field_name.as_ref();
let field = Field::new(field_name, array.data_type().clone(), nullable);
(field, array)
})
.unzip();
let schema = Arc::new(Schema::new(fields));
RecordBatch::try_new(schema, columns)
}
}
/// Options that control the behaviour used when creating a [`RecordBatch`].
#[derive(Debug)]
pub struct RecordBatchOptions {
/// Match field names of structs and lists. If set to `true`, the names must match.
pub match_field_names: bool,
}
impl Default for RecordBatchOptions {
fn default() -> Self {
Self {
match_field_names: true,
}
}
}
impl From<StructArray> for RecordBatch {
/// # Panics iff the null count of the array is not null.
fn from(array: StructArray) -> Self {
assert!(array.null_count() == 0);
let (fields, values, _) = array.into_data();
RecordBatch {
schema: Arc::new(Schema::new(fields)),
columns: values,
}
}
}
impl From<RecordBatch> for StructArray {
fn from(batch: RecordBatch) -> Self {
let (fields, values) = batch
.schema
.fields
.iter()
.zip(batch.columns.iter())
.map(|t| (t.0.clone(), t.1.clone()))
.unzip();
StructArray::from_data(DataType::Struct(fields), values, None)
}
}