Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Change ScalarValue::{List, LargeList, FixedSizedList} to take specific types rather than ArrayRef #8562

Merged
merged 18 commits into from
Jan 9, 2024
Merged

Change ScalarValue::{List, LargeList, FixedSizedList} to take specific types rather than ArrayRef #8562

Show file tree
Hide file tree
Changes from 5 commits
Commits
Show all changes
18 commits
Select commit Hold shift + click to select a range
4bc822b
Change ScalarValue::List type signature
Dec 15, 2023
6938bd6
Formatting/cleanup
Dec 15, 2023
a8585b6
Remove duplicate match statements
Dec 18, 2023
da1bf38
Add back scalar eq_array test for List
Dec 18, 2023
48bfc92
Formatting
Dec 18, 2023
bf1adaf
Reduce code duplication
Jan 6, 2024
5fe9329
Merge branch 'main' into scalarvalue-list-sig
rspears74 Jan 6, 2024
73d6776
Fix merge conflict
Jan 6, 2024
41f0d36
Fix post-merge compile errors
Jan 6, 2024
c1ec85e
Remove redundant partial_cmp implementation
alamb Jan 6, 2024
2fd67d9
improve
alamb Jan 6, 2024
84f3fed
Cargo fmt fix
Jan 6, 2024
7cf1511
Merge pull request #1 from alamb/alamb/less_copy3
rspears74 Jan 6, 2024
38488af
Reduce duplication in formatter
Jan 6, 2024
c212877
Reduce more duplication
Jan 6, 2024
588671c
Fix test error
Jan 6, 2024
2579749
Merge branch 'apache:main' into scalarvalue-list-sig
rspears74 Jan 7, 2024
713374c
Merge remote-tracking branch 'apache/main' into scalarvalue-list-sig
alamb Jan 9, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
291 changes: 202 additions & 89 deletions datafusion/common/src/scalar.rs
View file
Open in desktop

Large diffs are not rendered by default.

4 changes: 2 additions & 2 deletions datafusion/optimizer/src/simplify_expressions/expr_simplifier.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -27,7 +27,7 @@ use crate::simplify_expressions::regex::simplify_regex_expr;
use crate::simplify_expressions::SimplifyInfo;

use arrow::{
array::new_null_array,
array::{new_null_array, AsArray},
datatypes::{DataType, Field, Schema},
error::ArrowError,
record_batch::RecordBatch,
Expand Down Expand Up @@ -396,7 +396,7 @@ impl<'a> ConstEvaluator<'a> {
a.len()
)
} else if as_list_array(&a).is_ok() || as_large_list_array(&a).is_ok() {
Ok(ScalarValue::List(a))
Ok(ScalarValue::List(a.as_list().to_owned().into()))
} else {
// Non-ListArray
ScalarValue::try_from_array(&a, 0)
Expand Down
6 changes: 1 addition & 5 deletions datafusion/physical-expr/src/aggregate/array_agg_distinct.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -186,7 +186,6 @@ mod tests {
use arrow::array::{ArrayRef, Int32Array};
use arrow::datatypes::{DataType, Schema};
use arrow::record_batch::RecordBatch;
use arrow_array::cast::as_list_array;
use arrow_array::types::Int32Type;
use arrow_array::{Array, ListArray};
use arrow_buffer::OffsetBuffer;
Expand All @@ -196,10 +195,7 @@ mod tests {
// arrow::compute::sort cann't sort ListArray directly, so we need to sort the inner primitive array and wrap it back into ListArray.
fn sort_list_inner(arr: ScalarValue) -> ScalarValue {
let arr = match arr {
ScalarValue::List(arr) => {
let list_arr = as_list_array(&arr);
list_arr.value(0)
}
ScalarValue::List(arr) => arr.value(0),
_ => {
panic!("Expected ScalarValue::List, got {:?}", arr)
}
Expand Down
6 changes: 2 additions & 4 deletions datafusion/physical-expr/src/aggregate/tdigest.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -28,7 +28,6 @@
//! [Facebook's Folly TDigest]: https://github.com/facebook/folly/blob/main/folly/stats/TDigest.h

use arrow::datatypes::DataType;
use arrow_array::cast::as_list_array;
use arrow_array::types::Float64Type;
use datafusion_common::cast::as_primitive_array;
use datafusion_common::Result;
Expand Down Expand Up @@ -606,11 +605,10 @@ impl TDigest {

let centroids: Vec<_> = match &state[5] {
ScalarValue::List(arr) => {
let list_array = as_list_array(arr);
let arr = list_array.values();
let array = arr.values();

let f64arr =
as_primitive_array::<Float64Type>(arr).expect("expected f64 array");
as_primitive_array::<Float64Type>(array).expect("expected f64 array");
f64arr
.values()
.chunks(2)
Expand Down
3 changes: 2 additions & 1 deletion datafusion/physical-plan/src/values.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,6 +30,7 @@ use crate::{
use arrow::array::new_null_array;
use arrow::datatypes::SchemaRef;
use arrow::record_batch::RecordBatch;
use arrow_array::cast::AsArray;
use datafusion_common::{internal_err, plan_err, DataFusionError, Result, ScalarValue};
use datafusion_execution::TaskContext;

Expand Down Expand Up @@ -71,7 +72,7 @@ impl ValuesExec {
match r {
Ok(ColumnarValue::Scalar(scalar)) => Ok(scalar),
Ok(ColumnarValue::Array(a)) if a.len() == 1 => {
Ok(ScalarValue::List(a))
Ok(ScalarValue::List(a.as_list().to_owned().into()))
}
Ok(ColumnarValue::Array(a)) => {
plan_err!(
Expand Down
13 changes: 10 additions & 3 deletions datafusion/proto/src/logical_plan/from_proto.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -27,6 +27,7 @@ use crate::protobuf::{
OptimizedPhysicalPlanType, PlaceholderNode, RollupNode,
};
use arrow::{
array::AsArray,
buffer::Buffer,
datatypes::{
i256, DataType, Field, IntervalMonthDayNanoType, IntervalUnit, Schema, TimeUnit,
Expand Down Expand Up @@ -721,9 +722,15 @@ impl TryFrom<&protobuf::ScalarValue> for ScalarValue {
.map_err(|e| e.context("Decoding ScalarValue::List Value"))?;
let arr = record_batch.column(0);
match value {
Value::ListValue(_) => Self::List(arr.to_owned()),
Value::LargeListValue(_) => Self::LargeList(arr.to_owned()),
Value::FixedSizeListValue(_) => Self::FixedSizeList(arr.to_owned()),
Value::ListValue(_) => {
Self::List(arr.as_list::<i32>().to_owned().into())
}
Value::LargeListValue(_) => {
Self::LargeList(arr.as_list::<i64>().to_owned().into())
}
Value::FixedSizeListValue(_) => {
Self::FixedSizeList(arr.as_fixed_size_list().to_owned().into())
}
_ => unreachable!(),
}
}
Expand Down
98 changes: 76 additions & 22 deletions datafusion/proto/src/logical_plan/to_proto.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -32,6 +32,7 @@ use crate::protobuf::{
OptimizedLogicalPlanType, OptimizedPhysicalPlanType, PlaceholderNode, RollupNode,
};
use arrow::{
array::ArrayRef,
datatypes::{
DataType, Field, IntervalMonthDayNanoType, IntervalUnit, Schema, SchemaRef,
TimeUnit, UnionMode,
Expand Down Expand Up @@ -1159,13 +1160,11 @@ impl TryFrom<&ScalarValue> for protobuf::ScalarValue {
}
// ScalarValue::List and ScalarValue::FixedSizeList are serialized using
// Arrow IPC messages as a single column RecordBatch
ScalarValue::List(arr)
| ScalarValue::LargeList(arr)
| ScalarValue::FixedSizeList(arr) => {
ScalarValue::List(arr) => {
// Wrap in a "field_name" column
let batch = RecordBatch::try_from_iter(vec![(
"field_name",
arr.to_owned(),
arr.to_owned() as ArrayRef,
)])
.map_err(|e| {
Error::General( format!("Error creating temporary batch while encoding ScalarValue::List: {e}"))
Expand All @@ -1189,24 +1188,79 @@ impl TryFrom<&ScalarValue> for protobuf::ScalarValue {
schema: Some(schema),
};

match val {
ScalarValue::List(_) => Ok(protobuf::ScalarValue {
value: Some(protobuf::scalar_value::Value::ListValue(
scalar_list_value,
)),
}),
ScalarValue::LargeList(_) => Ok(protobuf::ScalarValue {
value: Some(protobuf::scalar_value::Value::LargeListValue(
scalar_list_value,
)),
}),
ScalarValue::FixedSizeList(_) => Ok(protobuf::ScalarValue {
value: Some(protobuf::scalar_value::Value::FixedSizeListValue(
scalar_list_value,
)),
}),
_ => unreachable!(),
}
Ok(protobuf::ScalarValue {
value: Some(protobuf::scalar_value::Value::ListValue(
scalar_list_value,
)),
})
}
ScalarValue::LargeList(arr) => {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Would it be possible to avoid the copy/paste in this method by making a function? Something like the following?

fn encode_scalar_list_value(array: &Array) -> Result<protobuf::ScalarListValue> {
...
}

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This one was easy enough to implement.

// Wrap in a "field_name" column
let batch = RecordBatch::try_from_iter(vec![(
"field_name",
arr.to_owned() as ArrayRef,
)])
.map_err(|e| {
Error::General( format!("Error creating temporary batch while encoding ScalarValue::List: {e}"))
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
Error::General( format!("Error creating temporary batch while encoding ScalarValue::List: {e}"))
Error::General( format!("Error creating temporary batch while encoding ScalarValue::LargeList: {e}"))

})?;

let gen = IpcDataGenerator {};
let mut dict_tracker = DictionaryTracker::new(false);
let (_, encoded_message) = gen
.encoded_batch(&batch, &mut dict_tracker, &Default::default())
.map_err(|e| {
Error::General(format!(
"Error encoding ScalarValue::List as IPC: {e}"
))
})?;

let schema: protobuf::Schema = batch.schema().try_into()?;

let scalar_list_value = protobuf::ScalarListValue {
ipc_message: encoded_message.ipc_message,
arrow_data: encoded_message.arrow_data,
schema: Some(schema),
};

Ok(protobuf::ScalarValue {
value: Some(protobuf::scalar_value::Value::LargeListValue(
scalar_list_value,
)),
})
}
ScalarValue::FixedSizeList(arr) => {
rspears74 marked this conversation as resolved.
Show resolved Hide resolved
// Wrap in a "field_name" column
let batch = RecordBatch::try_from_iter(vec![(
"field_name",
arr.to_owned() as ArrayRef,
)])
.map_err(|e| {
Error::General( format!("Error creating temporary batch while encoding ScalarValue::List: {e}"))
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
Error::General( format!("Error creating temporary batch while encoding ScalarValue::List: {e}"))
Error::General( format!("Error creating temporary batch while encoding ScalarValue::FixedSizeList: {e}"))

})?;

let gen = IpcDataGenerator {};
let mut dict_tracker = DictionaryTracker::new(false);
let (_, encoded_message) = gen
.encoded_batch(&batch, &mut dict_tracker, &Default::default())
.map_err(|e| {
Error::General(format!(
"Error encoding ScalarValue::List as IPC: {e}"
))
})?;

let schema: protobuf::Schema = batch.schema().try_into()?;

let scalar_list_value = protobuf::ScalarListValue {
ipc_message: encoded_message.ipc_message,
arrow_data: encoded_message.arrow_data,
schema: Some(schema),
};

Ok(protobuf::ScalarValue {
value: Some(protobuf::scalar_value::Value::FixedSizeListValue(
scalar_list_value,
)),
})
}
ScalarValue::Date32(val) => {
create_proto_scalar(val.as_ref(), &data_type, |s| Value::Date32Value(*s))
Expand Down