Skip to content

Fix AT correlation for join dimensions #13

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
nicosuave merged 2 commits into from
Jan 9, 2026
Merged

Fix AT correlation for join dimensions #13

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
680e25a
Fix AT correlation with joins
nicosuave Jan 9, 2026
eb83918
Preserve alias-based group by dims
nicosuave Jan 9, 2026
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
62 changes: 62 additions & 0 deletions test/sql/measures.test
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -761,6 +761,34 @@ FROM daily_orders_v
2 320.0 840.0
3 270.0 840.0

# =============================================================================
# Test: GROUP BY alias for expression dimension
# =============================================================================

statement ok
CREATE TABLE monthly_sales (order_date DATE, region TEXT, amount DOUBLE);

statement ok
INSERT INTO monthly_sales VALUES
('2023-01-05', 'US', 100), ('2023-01-12', 'EU', 50),
('2023-02-03', 'US', 200), ('2023-02-20', 'EU', 20);

statement ok
CREATE VIEW monthly_sales_v AS
SELECT DATE_TRUNC('month', order_date) AS month, region, SUM(amount) AS MEASURE revenue
FROM monthly_sales
GROUP BY DATE_TRUNC('month', order_date), region;

query IIRR rowsort
SEMANTIC SELECT month, region, AGGREGATE(revenue), AGGREGATE(revenue) AT (ALL region) AS month_total
FROM monthly_sales_v
;
----
2023-01-01 EU 50.0 150.0
2023-01-01 US 100.0 150.0
2023-02-01 EU 20.0 220.0
2023-02-01 US 200.0 220.0

# =============================================================================
# Test: Multi-fact JOINs (wide tables)
# =============================================================================
Expand Down Expand Up @@ -818,6 +846,40 @@ FROM fact_orders_v o JOIN fact_returns_v r ON o.year = r.year AND o.region = r.r
2023 EU 75.0 225.0
2023 US 150.0 225.0

# =============================================================================
# Test: JOIN with extra dimension from second table
# =============================================================================

statement ok
CREATE TABLE salesdetails (year INT, region TEXT, product TEXT, amount DOUBLE);

statement ok
INSERT INTO salesdetails VALUES
(2022, 'US', 'Shoes', 2), (2022, 'US', 'Cars', 1),
(2022, 'EU', 'Shoes', 3),
(2023, 'US', 'Shoes', 4), (2023, 'US', 'Cars', 2),
(2023, 'EU', 'Cars', 5);

statement ok
CREATE VIEW salesdetails_v AS
SELECT year, region, product, SUM(amount) AS MEASURE quantity
FROM salesdetails;

query IIIRRR rowsort
SEMANTIC SELECT s.year, s.region, sd.product,
AGGREGATE(revenue) AS year_sales_revenue,
AGGREGATE(revenue) AT (ALL year) AS region_total,
AGGREGATE(quantity) AS product_qty
FROM sales_v s JOIN salesdetails_v sd ON s.year = sd.year AND s.region = sd.region
;
----
2022 EU Shoes 50.0 125.0 3.0
2022 US Cars 100.0 250.0 1.0
2022 US Shoes 100.0 250.0 2.0
2023 EU Cars 75.0 125.0 5.0
2023 US Cars 150.0 250.0 2.0
2023 US Shoes 150.0 250.0 4.0

# =============================================================================
# Test: SET reaches beyond WHERE clause (paper semantics)
# Per paper: SET should evaluate over data removed by outer WHERE clause
Expand Down
70 changes: 63 additions & 7 deletions yardstick-rs/src/sql/measures.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,7 +7,7 @@
//!
//! Reference: https://arxiv.org/abs/2406.00251

use std::collections::HashMap;
use std::collections::{HashMap, HashSet};
use std::sync::Mutex;

use nom::{
Expand Down Expand Up @@ -1051,6 +1051,52 @@ fn group_by_matches_view(outer_cols: &[String], view_cols: &[String]) -> bool {
!outer_set.is_empty() && outer_set == view_set
}

fn filter_group_by_cols_for_measure(
outer_cols: &[String],
view_cols: &[String],
dimension_exprs: &HashMap<String, String>,
) -> Vec<String> {
if view_cols.is_empty() {
return outer_cols.to_vec();
}

let view_set: HashSet<String> = view_cols
.iter()
.map(|col| normalize_group_by_col(col))
.collect();

let mut alias_expr_norms: Vec<(String, String)> = Vec::new();
alias_expr_norms.reserve(dimension_exprs.len());
for (alias, expr) in dimension_exprs.iter() {
alias_expr_norms.push((
normalize_group_by_col(alias),
normalize_group_by_col(expr),
));
}

outer_cols
.iter()
.filter(|col| {
let normalized_outer = normalize_group_by_col(col);
if view_set.contains(&normalized_outer) {
return true;
}

if let Some(expr) = dimension_exprs.get(&normalized_outer) {
let normalized_expr = normalize_group_by_col(expr);
if view_set.contains(&normalized_expr) {
return true;
}
}

alias_expr_norms.iter().any(|(alias_norm, expr_norm)| {
expr_norm == &normalized_outer && view_set.contains(alias_norm)
})
})
.cloned()
.collect()
}

fn can_use_view_measure_directly(resolved: &ResolvedMeasure, outer_group_by: &[String]) -> bool {
group_by_matches_view(outer_group_by, &resolved.view_group_by_cols)
}
Expand Down Expand Up @@ -3416,6 +3462,11 @@ pub fn expand_aggregate_with_at(sql: &str) -> AggregateExpandResult {
for (measure_name, modifiers, start, end) in patterns {
// Look up which view contains this measure (for JOIN support)
let resolved = resolve_measure_source(&measure_name, &primary_table_name);
let measure_group_by_cols = filter_group_by_cols_for_measure(
&group_by_cols,
&resolved.view_group_by_cols,
&resolved.dimension_exprs,
);

// Non-decomposable measures are recomputed from base rows (including AT modifiers)

Expand Down Expand Up @@ -3452,7 +3503,7 @@ pub fn expand_aggregate_with_at(sql: &str) -> AggregateExpandResult {
&resolved.source_view,
outer_alias_ref,
outer_where_ref,
&group_by_cols,
&measure_group_by_cols,
)
} else if !resolved.is_decomposable {
let outer_ref_for_non_decomp =
Expand Down Expand Up @@ -3480,7 +3531,7 @@ pub fn expand_aggregate_with_at(sql: &str) -> AggregateExpandResult {
&base_relation_sql,
outer_ref_for_non_decomp,
outer_where_ref,
&group_by_cols,
&measure_group_by_cols,
&modifiers,
&resolved.dimension_exprs,
&format!("_nd_{join_counter}"),
Expand All @@ -3496,7 +3547,7 @@ pub fn expand_aggregate_with_at(sql: &str) -> AggregateExpandResult {
&base_relation_sql,
outer_ref_for_non_decomp,
outer_where_ref,
&group_by_cols,
&measure_group_by_cols,
&modifiers,
&resolved.dimension_exprs,
),
Expand All @@ -3517,7 +3568,7 @@ pub fn expand_aggregate_with_at(sql: &str) -> AggregateExpandResult {
&resolved.source_view,
outer_alias_ref,
outer_where_ref,
&group_by_cols,
&measure_group_by_cols,
)
};
result_sql = format!("{}{}{}", &result_sql[..start], expanded, &result_sql[end..]);
Expand All @@ -3529,6 +3580,11 @@ pub fn expand_aggregate_with_at(sql: &str) -> AggregateExpandResult {

for (measure_name, start, end) in plain_calls {
let resolved = resolve_measure_source(&measure_name, &primary_table_name);
let measure_group_by_cols = filter_group_by_cols_for_measure(
&group_by_cols,
&resolved.view_group_by_cols,
&resolved.dimension_exprs,
);


// For derived measures, use the expanded expression; otherwise use AGG(measure_name)
Expand Down Expand Up @@ -3558,7 +3614,7 @@ pub fn expand_aggregate_with_at(sql: &str) -> AggregateExpandResult {
&base_relation_sql,
outer_ref_for_non_decomp,
outer_where_ref,
&group_by_cols,
&measure_group_by_cols,
&[], // No modifiers for plain AGGREGATE()
&resolved.dimension_exprs,
&format!("_nd_{join_counter}"),
Expand All @@ -3574,7 +3630,7 @@ pub fn expand_aggregate_with_at(sql: &str) -> AggregateExpandResult {
&base_relation_sql,
outer_ref_for_non_decomp,
outer_where_ref,
&group_by_cols,
&measure_group_by_cols,
&[], // No modifiers for plain AGGREGATE()
&resolved.dimension_exprs,
),
Expand Down
Loading