Skip to content

fix: Ignore nullability of list elements when consuming Substrait #10874

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

Merged
merged 3 commits into from
Jun 12, 2024
Merged

fix: Ignore nullability of list elements when consuming Substrait #10874

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
f1d124b
Ignore nullability of list elements when consuming Substrait
Jun 11, 2024
e873b5e
use `Field::new_list_field` in `array_into_(large_)list_array`
Jun 11, 2024
a25abd8
add a test for non-nullable lists
Jun 11, 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
14 changes: 7 additions & 7 deletions datafusion/common/src/utils/mod.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -354,7 +354,7 @@ pub fn longest_consecutive_prefix<T: Borrow<usize>>(
pub fn array_into_list_array(arr: ArrayRef) -> ListArray {
let offsets = OffsetBuffer::from_lengths([arr.len()]);
ListArray::new(
Arc::new(Field::new("item", arr.data_type().to_owned(), true)),
Arc::new(Field::new_list_field(arr.data_type().to_owned(), true)),
Copy link
Contributor Author

@Blizzara Blizzara Jun 11, 2024

Choose a reason for hiding this comment

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

this is no-op change, just seems nicer to use the new_list_field given it exists (it sets the name as "item" anyways)

offsets,
arr,
None,
Expand All @@ -366,7 +366,7 @@ pub fn array_into_list_array(arr: ArrayRef) -> ListArray {
pub fn array_into_large_list_array(arr: ArrayRef) -> LargeListArray {
let offsets = OffsetBuffer::from_lengths([arr.len()]);
LargeListArray::new(
Arc::new(Field::new("item", arr.data_type().to_owned(), true)),
Arc::new(Field::new_list_field(arr.data_type().to_owned(), true)),
Copy link
Contributor

@alamb alamb Jun 11, 2024

Choose a reason for hiding this comment

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

👍

offsets,
arr,
None,
Expand All @@ -379,7 +379,7 @@ pub fn array_into_fixed_size_list_array(
) -> FixedSizeListArray {
let list_size = list_size as i32;
FixedSizeListArray::new(
Arc::new(Field::new("item", arr.data_type().to_owned(), true)),
Arc::new(Field::new_list_field(arr.data_type().to_owned(), true)),
list_size,
arr,
None,
Expand Down Expand Up @@ -420,7 +420,7 @@ pub fn arrays_into_list_array(
let data_type = arr[0].data_type().to_owned();
let values = arr.iter().map(|x| x.as_ref()).collect::<Vec<_>>();
Ok(ListArray::new(
Arc::new(Field::new("item", data_type, true)),
Arc::new(Field::new_list_field(data_type, true)),
OffsetBuffer::from_lengths(lens),
arrow::compute::concat(values.as_slice())?,
None,
Expand All @@ -435,7 +435,7 @@ pub fn arrays_into_list_array(
/// use datafusion_common::utils::base_type;
/// use std::sync::Arc;
///
/// let data_type = DataType::List(Arc::new(Field::new("item", DataType::Int32, true)));
/// let data_type = DataType::List(Arc::new(Field::new_list_field(DataType::Int32, true)));
/// assert_eq!(base_type(&data_type), DataType::Int32);
///
/// let data_type = DataType::Int32;
Expand All @@ -458,10 +458,10 @@ pub fn base_type(data_type: &DataType) -> DataType {
/// use datafusion_common::utils::coerced_type_with_base_type_only;
/// use std::sync::Arc;
///
/// let data_type = DataType::List(Arc::new(Field::new("item", DataType::Int32, true)));
/// let data_type = DataType::List(Arc::new(Field::new_list_field(DataType::Int32, true)));
/// let base_type = DataType::Float64;
/// let coerced_type = coerced_type_with_base_type_only(&data_type, &base_type);
/// assert_eq!(coerced_type, DataType::List(Arc::new(Field::new("item", DataType::Float64, true))));
/// assert_eq!(coerced_type, DataType::List(Arc::new(Field::new_list_field(DataType::Float64, true))));
pub fn coerced_type_with_base_type_only(
data_type: &DataType,
base_type: &DataType,
Expand Down
4 changes: 3 additions & 1 deletion datafusion/substrait/src/logical_plan/consumer.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1272,7 +1272,9 @@ fn from_substrait_type(
})?;
let field = Arc::new(Field::new_list_field(
from_substrait_type(inner_type, dfs_names, name_idx)?,
is_substrait_type_nullable(inner_type)?,
// We ignore Substrait's nullability here to match to_substrait_literal
Copy link
Contributor

@alamb alamb Jun 11, 2024

Choose a reason for hiding this comment

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

👍

// which always creates nullable lists
true,
));
match list.type_variation_reference {
DEFAULT_CONTAINER_TYPE_VARIATION_REF => Ok(DataType::List(field)),
Expand Down
14 changes: 6 additions & 8 deletions datafusion/substrait/src/logical_plan/producer.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2309,14 +2309,12 @@ mod test {
round_trip_type(DataType::Decimal128(10, 2))?;
round_trip_type(DataType::Decimal256(30, 2))?;

for nullable in [true, false] {
round_trip_type(DataType::List(
Field::new_list_field(DataType::Int32, nullable).into(),
))?;
round_trip_type(DataType::LargeList(
Field::new_list_field(DataType::Int32, nullable).into(),
))?;
}
round_trip_type(DataType::List(
Copy link
Contributor

@alamb alamb Jun 11, 2024

Choose a reason for hiding this comment

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

Should we also add a test here showing that the List becomes nullable after roundtrip? That might additionally document that this is an intended behavior rather than a bug.

I do see you have added a comment which I think is probably enought

Copy link
Contributor Author

@Blizzara Blizzara Jun 11, 2024 •
edited
Loading

Choose a reason for hiding this comment

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

I don't think we need to test the null coercion specifically, since that's not necessarily the ultimate desired behavior, just what makes sense at this time. But I added a test case to confirm that we can now read plans with non-nullable lists: a25abd8

Field::new_list_field(DataType::Int32, true).into(),
))?;
round_trip_type(DataType::LargeList(
Field::new_list_field(DataType::Int32, true).into(),
))?;

round_trip_type(DataType::Struct(
vec![
Expand Down
32 changes: 27 additions & 5 deletions datafusion/substrait/tests/cases/logical_plans.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,6 +20,7 @@
#[cfg(test)]
mod tests {
use datafusion::common::Result;
use datafusion::dataframe::DataFrame;
use datafusion::prelude::{CsvReadOptions, SessionContext};
use datafusion_substrait::logical_plan::consumer::from_substrait_plan;
use std::fs::File;
Expand All @@ -38,11 +39,7 @@ mod tests {

// File generated with substrait-java's Isthmus:
// ./isthmus-cli/build/graal/isthmus "select not d from data" -c "create table data (d boolean)"
let path = "tests/testdata/select_not_bool.substrait.json";
let proto = serde_json::from_reader::<_, Plan>(BufReader::new(
File::open(path).expect("file not found"),
))
.expect("failed to parse json");
let proto = read_json("tests/testdata/select_not_bool.substrait.json");

let plan = from_substrait_plan(&ctx, &proto).await?;

Expand All @@ -54,6 +51,31 @@ mod tests {
Ok(())
}

#[tokio::test]
async fn non_nullable_lists() -> Result<()> {
// DataFusion's Substrait consumer treats all lists as nullable, even if the Substrait plan specifies them as non-nullable.
// That's because implementing the non-nullability consistently is non-trivial.
// This test confirms that reading a plan with non-nullable lists works as expected.
let ctx = create_context().await?;
let proto = read_json("tests/testdata/non_nullable_lists.substrait.json");

let plan = from_substrait_plan(&ctx, &proto).await?;

assert_eq!(format!("{:?}", &plan), "Values: (List([1, 2]))");

// Need to trigger execution to ensure that Arrow has validated the plan
DataFrame::new(ctx.state(), plan).show().await?;

Ok(())
}

fn read_json(path: &str) -> Plan {
serde_json::from_reader::<_, Plan>(BufReader::new(
File::open(path).expect("file not found"),
))
.expect("failed to parse json")
}

async fn create_context() -> datafusion::common::Result<SessionContext> {
let ctx = SessionContext::new();
ctx.register_csv("DATA", "tests/testdata/data.csv", CsvReadOptions::new())
Expand Down
71 changes: 71 additions & 0 deletions datafusion/substrait/tests/testdata/non_nullable_lists.substrait.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,71 @@
{
"extensionUris": [],
"extensions": [],
"relations": [
{
"root": {
"input": {
"read": {
"common": {
"direct": {
}
},
"baseSchema": {
"names": [
"col"
],
"struct": {
"types": [
{
"list": {
"type": {
"i32": {
"typeVariationReference": 0,
"nullability": "NULLABILITY_REQUIRED"
}
},
"typeVariationReference": 0,
"nullability": "NULLABILITY_REQUIRED"
}
}
],
"typeVariationReference": 0,
"nullability": "NULLABILITY_REQUIRED"
}
},
"virtualTable": {
"values": [
{
"fields": [
{
"list": {
"values": [
{
"i32": 1,
"nullable": false,
"typeVariationReference": 0
},
{
"i32": 2,
"nullable": false,
"typeVariationReference": 0
}
]
},
"nullable": false,
"typeVariationReference": 0
}
]
}
]
}
}
},
"names": [
"col"
]
}
}
],
"expectedTypeUrls": []
}