Skip to content
Draft
Show file tree
Hide file tree
Changes from all commits
Commits
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
313 changes: 312 additions & 1 deletion datafusion/core/tests/sql/joins.rs
Original file line number Diff line number Diff line change
Expand Up @@ -17,10 +17,12 @@

use insta::assert_snapshot;

use datafusion::assert_batches_eq;
use datafusion::catalog::MemTable;
use datafusion::datasource::stream::{FileStreamProvider, StreamConfig, StreamTable};
use datafusion::physical_plan::joins::AsOfJoinExec;
use datafusion::physical_plan::{Distribution, ExecutionPlanProperties};
use datafusion::test_util::register_unbounded_file_with_ordering;
use datafusion::{assert_batches_eq, assert_batches_sorted_eq};
use datafusion_sql::unparser::plan_to_sql;

use super::*;
Expand Down Expand Up @@ -297,3 +299,312 @@ async fn unparse_cross_join() -> Result<()> {

Ok(())
}

fn register_asof_test_tables(ctx: &SessionContext) -> Result<()> {
let trades_schema = Arc::new(Schema::new(vec![
Field::new("symbol", DataType::Utf8, true),
Field::new("ts", DataType::Int64, true),
Field::new("trade_id", DataType::Int32, false),
]));
let trades = vec![
RecordBatch::try_new(
Arc::clone(&trades_schema),
vec![
Arc::new(StringArray::from(vec![Some("A"), Some("B"), None])),
Arc::new(Int64Array::from(vec![Some(7), Some(2), Some(3)])),
Arc::new(Int32Array::from(vec![3, 4, 6])),
],
)?,
RecordBatch::try_new(
Arc::clone(&trades_schema),
vec![
Arc::new(StringArray::from(vec![Some("A"), Some("A"), Some("B")])),
Arc::new(Int64Array::from(vec![Some(1), Some(4), Some(8)])),
Arc::new(Int32Array::from(vec![1, 2, 5])),
],
)?,
];
ctx.register_table(
"trades",
Arc::new(MemTable::try_new(
trades_schema,
trades.into_iter().map(|batch| vec![batch]).collect(),
)?),
)?;

let prices_schema = Arc::new(Schema::new(vec![
Field::new("symbol", DataType::Utf8, true),
Field::new("ts", DataType::Int64, true),
Field::new("price", DataType::Int32, false),
]));
let prices = vec![
RecordBatch::try_new(
Arc::clone(&prices_schema),
vec![
Arc::new(StringArray::from(vec![Some("A"), Some("B"), None])),
Arc::new(Int64Array::from(vec![Some(6), Some(1), Some(2)])),
Arc::new(Int32Array::from(vec![60, 101, 999])),
],
)?,
RecordBatch::try_new(
Arc::clone(&prices_schema),
vec![
Arc::new(StringArray::from(vec![Some("A"), Some("A"), Some("B")])),
Arc::new(Int64Array::from(vec![Some(2), Some(4), Some(6)])),
Arc::new(Int32Array::from(vec![20, 40, 106])),
],
)?,
];
ctx.register_table(
"prices",
Arc::new(MemTable::try_new(
prices_schema,
prices.into_iter().map(|batch| vec![batch]).collect(),
)?),
)?;
Ok(())
}

fn find_asof_exec(plan: &Arc<dyn ExecutionPlan>) -> Option<Arc<dyn ExecutionPlan>> {
if plan.downcast_ref::<AsOfJoinExec>().is_some() {
return Some(Arc::clone(plan));
}
plan.children().into_iter().find_map(find_asof_exec)
}

#[tokio::test]
async fn asof_join_all_match_directions_across_batches() -> Result<()> {
let config = SessionConfig::new()
.with_batch_size(2)
.with_target_partitions(2);
let ctx = SessionContext::new_with_config(config);
register_asof_test_tables(&ctx)?;

for (op, expected) in [
(
">=",
[
"+----------+-------+",
"| trade_id | price |",
"+----------+-------+",
"| 1 | |",
"| 2 | 40 |",
"| 3 | 60 |",
"| 4 | 101 |",
"| 5 | 106 |",
"| 6 | |",
"+----------+-------+",
],
),
(
">",
[
"+----------+-------+",
"| trade_id | price |",
"+----------+-------+",
"| 1 | |",
"| 2 | 20 |",
"| 3 | 60 |",
"| 4 | 101 |",
"| 5 | 106 |",
"| 6 | |",
"+----------+-------+",
],
),
(
"<=",
[
"+----------+-------+",
"| trade_id | price |",
"+----------+-------+",
"| 1 | 20 |",
"| 2 | 40 |",
"| 3 | |",
"| 4 | 106 |",
"| 5 | |",
"| 6 | |",
"+----------+-------+",
],
),
(
"<",
[
"+----------+-------+",
"| trade_id | price |",
"+----------+-------+",
"| 1 | 20 |",
"| 2 | 60 |",
"| 3 | |",
"| 4 | 106 |",
"| 5 | |",
"| 6 | |",
"+----------+-------+",
],
),
] {
let batches = ctx
.sql(&format!(
"SELECT t.trade_id, p.price FROM trades t \
ASOF JOIN prices p MATCH_CONDITION (t.ts {op} p.ts) \
ON t.symbol = p.symbol ORDER BY t.trade_id"
))
.await?
.collect()
.await?;
assert_batches_eq!(expected, &batches);
}
Ok(())
}

#[tokio::test]
async fn asof_join_broadcasts_multi_partition_right_input() -> Result<()> {
let config = SessionConfig::new().with_target_partitions(4);
let ctx = SessionContext::new_with_config(config);
register_asof_test_tables(&ctx)?;
let df = ctx
.sql(
"SELECT t.trade_id, p.price FROM trades t ASOF JOIN \
(SELECT ts, price FROM prices WHERE symbol = 'A') p \
MATCH_CONDITION (t.ts >= p.ts)",
)
.await?;
let sql = plan_to_sql(df.logical_plan())?.to_string();
assert_contains!(sql.as_str(), "ASOF JOIN");
assert!(!sql.contains(" ON "), "unexpected equality clause: {sql}");
ctx.sql(&sql).await?;
let plan = df.create_physical_plan().await?;
let asof = find_asof_exec(&plan).expect("physical ASOF join must be present");
let output_partitions = asof.output_partitioning().partition_count();
assert_eq!(
output_partitions,
asof.children()[0].output_partitioning().partition_count()
);
assert!(
output_partitions > 1,
"ASOF join did not preserve left-side parallelism"
);
assert_eq!(
asof.children()[1].output_partitioning().partition_count(),
1
);
let right_plan = displayable(asof.children()[1].as_ref())
.indent(true)
.to_string();
assert_contains!(right_plan.as_str(), "SortPreservingMergeExec");
assert_contains!(right_plan.as_str(), "DataSourceExec: partitions=2");
assert!(asof.output_ordering().is_some());
assert!(matches!(
&asof.input_distribution_requirements().into_per_child()[..],
[
Distribution::UnspecifiedDistribution,
Distribution::SinglePartition
]
));
let batches = collect(plan, ctx.task_ctx()).await?;
assert_batches_sorted_eq!(
[
"+----------+-------+",
"| trade_id | price |",
"+----------+-------+",
"| 1 | |",
"| 2 | 40 |",
"| 3 | 60 |",
"| 4 | 20 |",
"| 5 | 60 |",
"| 6 | 20 |",
"+----------+-------+",
],
&batches
);
Ok(())
}

#[tokio::test]
async fn asof_join_rejects_unbounded_inputs_during_physical_planning() -> Result<()> {
let ctx = SessionContext::new();
let tmp_dir = TempDir::new()?;
let schema = Arc::new(Schema::new(vec![
Field::new("symbol", DataType::UInt32, false),
Field::new("ts", DataType::UInt32, false),
]));
let ordering = vec![vec![
col("symbol").sort(true, true),
col("ts").sort(true, true),
]];
for table in ["left_stream", "right_stream"] {
let path = tmp_dir.path().join(format!("{table}.csv"));
File::create(&path)?;
register_unbounded_file_with_ordering(
&ctx,
Arc::clone(&schema),
&path,
table,
ordering.clone(),
)?;
}
let error = ctx
.sql(
"SELECT * FROM left_stream l ASOF JOIN right_stream r \
MATCH_CONDITION (l.ts >= r.ts) ON l.symbol = r.symbol",
)
.await?
.create_physical_plan()
.await
.expect_err("ASOF physical planning must reject unbounded inputs");
assert_contains!(error.to_string(), "AsOfJoinExec requires bounded inputs");
Ok(())
}

#[tokio::test]
async fn asof_join_using_unparser_round_trips() -> Result<()> {
let ctx = SessionContext::new();
register_asof_test_tables(&ctx)?;
let df = ctx
.sql(
"SELECT * FROM trades t ASOF JOIN prices p \
MATCH_CONDITION (t.ts >= p.ts) USING (symbol)",
)
.await?;
let sql = plan_to_sql(df.logical_plan())?.to_string();
assert!(sql.contains("ASOF JOIN"));
assert!(sql.contains("MATCH_CONDITION"));
assert!(sql.contains("USING(symbol)"), "unexpected SQL: {sql}");
ctx.sql(&sql).await?;
Ok(())
}

#[tokio::test]
async fn asof_join_unparser_preserves_right_preselection() -> Result<()> {
let ctx = SessionContext::new();
register_asof_test_tables(&ctx)?;
for query in [
"SELECT t.trade_id, p.price FROM trades t \
ASOF JOIN (SELECT * FROM prices WHERE price < 100) p \
MATCH_CONDITION (t.ts >= p.ts) ON t.symbol = p.symbol \
ORDER BY t.trade_id",
"SELECT * FROM trades t \
ASOF JOIN (SELECT * FROM prices WHERE price < 100) p \
MATCH_CONDITION (t.ts >= p.ts) USING (symbol) \
ORDER BY t.trade_id",
"SELECT t.trade_id, p.price FROM trades t \
JOIN prices q ON t.symbol = q.symbol AND t.ts = q.ts \
ASOF JOIN prices p MATCH_CONDITION (t.ts >= p.ts) \
ON q.symbol = p.symbol ORDER BY t.trade_id",
"SELECT t.trade_id, q.trade_id FROM trades t \
ASOF JOIN (prices p JOIN trades q \
ON p.symbol = q.symbol AND p.ts = q.ts) \
MATCH_CONDITION (t.ts >= q.ts) ON t.symbol = p.symbol \
ORDER BY t.trade_id",
] {
let expected = ctx.sql(query).await?.collect().await?;
let plan = ctx.sql(query).await?.into_optimized_plan()?;
let sql = plan_to_sql(&plan)?.to_string();
let actual = ctx.sql(&sql).await?.collect().await?;
assert_eq!(
datafusion_common::test_util::batches_to_string(&expected),
datafusion_common::test_util::batches_to_string(&actual),
"unparsed SQL changed ASOF candidate preselection: {sql}"
);
}
Ok(())
}
Loading