✔ PASS Search & Retrieval task_066_Q66_alphanumeric_sku_normalization Scaled: 1M+ Records Ordered: No

Alphanumeric SKU Matching with Punctuation & Case Invariance

Candidate Latency (1M Rows)
321.96 ms
(1.0x faster than golden)
Golden Latency (1M Rows)
325.47 ms
DuckDB in-memory engine
Underlying Data Volume
1,000,000 rows
42.0 MiB across 1 table
Evaluated Model
gemini-3.8-flash-low
Gen Latency: 15.6s
✔ Correctness Verified on 1M+ Records
Candidate SQL output matches ground truth golden query across all multiset rows and numeric tolerances (tol=1e-3). Zero data skew or anti-patterns detected on full dataset scale.

1. Analytical Significance & What It Means

Scenario & Frontier Trap
Business Scenario & Problem

Customers search for model numbers in messy formats: 'wh-1000xm5', 'WH1000-XM5', 'wh 1000 xm5'. Robust search normalizes punctuation and whitespace before matching.

⚡ The Frontier Trap (Why LLMs Fail)

Exact string equality misses 90% of technical product searches due to hyphens and spaces.

Prompt / User Request Given to Model
Customers search for model numbers in messy formats: 'wh-1000xm5', 'WH1000-XM5', 'wh 1000 xm5'. Robust search normalizes punctuation and whitespace before matching.

Execute the required analytical SQL calculation in DuckDB.
Expected output columns: sku_id, raw_sku, product_title.
Row order does not matter.

2. Underlying Tables, Row Counts & Data Volume

1 Tables (1,000,000 rows • 42.0 MiB)

Production-scale fixtures loaded into the isolated DuckDB sandbox with at least 1,000,000 records per table:

📊 inventory_skus 1,000,000 rows 42.0 MiB Data Volume
sku_id (INTEGER) raw_sku (VARCHAR) product_title (VARCHAR)
Data Preview (showing sample rows from 1,000,000 total rows):
# sku_id raw_sku product_title
1 1 WH-1000XM5/B Sony Premium Wireless Headphones Black
2 2 AIR-MAX-90-W Nike Air Max 90 White Sneaker
3 2 WH-1000XM5/B Sony Premium Wireless Headphones Black
4 3 AIR-MAX-90-W Nike Air Max 90 White Sneaker
5 4 WH-1000XM5/B Sony Premium Wireless Headphones Black
6 5 AIR-MAX-90-W Nike Air Max 90 White Sneaker
7 6 WH-1000XM5/B Sony Premium Wireless Headphones Black
8 7 AIR-MAX-90-W Nike Air Max 90 White Sneaker
9 8 WH-1000XM5/B Sony Premium Wireless Headphones Black
10 9 AIR-MAX-90-W Nike Air Max 90 White Sneaker
11 10 WH-1000XM5/B Sony Premium Wireless Headphones Black
12 11 AIR-MAX-90-W Nike Air Max 90 White Sneaker
13 12 WH-1000XM5/B Sony Premium Wireless Headphones Black
14 13 AIR-MAX-90-W Nike Air Max 90 White Sneaker
15 14 WH-1000XM5/B Sony Premium Wireless Headphones Black

3. Graphical Query Execution Plan & Cost Bottlenecks

Physical Operator Tree (DuckDB 2.0 Profiler)

Visual flowchart of the physical execution pipeline. Operators with high relative costs (≥ 20% of query runtime or heavy latency) are visually highlighted:

Pipeline Latency Distribution (Total Plan Time: 603.36 ms) Nodes consuming ≥ 20% of query runtime are highlighted with glowing borders
📐 PROJECTION 0.0%
Latency: 0.29 ms
Rows In/Out: 500,000
Data Vol: 17.17 MB
📁 Projections: ['sku_id', 'raw_sku...
🔍 TABLE_SCAN 🔥 100.0% Cost
Latency: 603.08 ms
Rows In/Out: 500,000
Data Vol: 17.17 MB
📁 memory.main.inventory_skus
⚠️ Heavy Execution Bottleneck (100.0% of query runtime)
Operator Cost % Latency Processed Rows Data Volume Details / Table Status
TABLE_SCAN 100.0% 603.08 ms 500,000 17.17 MB memory.main.inventory_skus 🔥 BOTTLENECK
PROJECTION 0.0% 0.29 ms 500,000 17.17 MB Projections: ['sku_id', 'raw_sku', 'product_titl Normal

4. Queries Executed (Candidate vs Golden Ground Truth)

Beautified & Syntax Highlighted
Candidate SQL Executed by gemini-3.8-flash-low (Runtime on 1M Rows: 321.96 ms)
SELECT sku_id,
       raw_sku,
       product_title
FROM inventory_skus
WHERE lower(regexp_replace(raw_sku, '[^a-zA-Z0-9]', '', 'g')) = lower(regexp_replace('wh 1000-xm5/b', '[^a-zA-Z0-9]', '', 'g'));
Golden Reference Ground Truth SQL (Runtime on 1M Rows: 325.47 ms)
-- Cleanse search query 'wh1000xm5b' by stripping non-alphanumeric characters

SELECT sku_id,
       raw_sku,
       product_title
FROM inventory_skus
WHERE lower(regexp_replace(raw_sku, '[^a-zA-Z0-9]', '', 'g')) = lower(regexp_replace('wh 1000-xm5/b', '[^a-zA-Z0-9]', '', 'g'));
DuckDB Fixture Setup DDL & Data Population
CREATE TABLE inventory_skus (sku_id INT, raw_sku VARCHAR, product_title VARCHAR);


INSERT INTO inventory_skus
VALUES (1, 'WH-1000XM5/B', 'Sony Premium Wireless Headphones Black'),
       (2, 'AIR-MAX-90-W', 'Nike Air Max 90 White Sneaker');

5. Query Results & Expected Ground Truth

Tabular Comparison
Actual Candidate Output (500,000 rows × 3 cols)
# sku_id raw_sku product_title
1 1 WH-1000XM5/B Sony Premium Wireless Headphones Black
2 2 WH-1000XM5/B Sony Premium Wireless Headphones Black
3 4 WH-1000XM5/B Sony Premium Wireless Headphones Black
4 6 WH-1000XM5/B Sony Premium Wireless Headphones Black
5 8 WH-1000XM5/B Sony Premium Wireless Headphones Black
6 10 WH-1000XM5/B Sony Premium Wireless Headphones Black
7 12 WH-1000XM5/B Sony Premium Wireless Headphones Black
8 14 WH-1000XM5/B Sony Premium Wireless Headphones Black
9 16 WH-1000XM5/B Sony Premium Wireless Headphones Black
10 18 WH-1000XM5/B Sony Premium Wireless Headphones Black
11 20 WH-1000XM5/B Sony Premium Wireless Headphones Black
12 22 WH-1000XM5/B Sony Premium Wireless Headphones Black
13 24 WH-1000XM5/B Sony Premium Wireless Headphones Black
14 26 WH-1000XM5/B Sony Premium Wireless Headphones Black
15 28 WH-1000XM5/B Sony Premium Wireless Headphones Black
16 30 WH-1000XM5/B Sony Premium Wireless Headphones Black
17 32 WH-1000XM5/B Sony Premium Wireless Headphones Black
18 34 WH-1000XM5/B Sony Premium Wireless Headphones Black
19 36 WH-1000XM5/B Sony Premium Wireless Headphones Black
20 38 WH-1000XM5/B Sony Premium Wireless Headphones Black
21 40 WH-1000XM5/B Sony Premium Wireless Headphones Black
22 42 WH-1000XM5/B Sony Premium Wireless Headphones Black
23 44 WH-1000XM5/B Sony Premium Wireless Headphones Black
24 46 WH-1000XM5/B Sony Premium Wireless Headphones Black
25 48 WH-1000XM5/B Sony Premium Wireless Headphones Black
Showing first 25 of 500,000 total rows.
Expected Golden Output (500,000 rows × 3 cols)
# sku_id raw_sku product_title
1 1 WH-1000XM5/B Sony Premium Wireless Headphones Black
2 2 WH-1000XM5/B Sony Premium Wireless Headphones Black
3 4 WH-1000XM5/B Sony Premium Wireless Headphones Black
4 6 WH-1000XM5/B Sony Premium Wireless Headphones Black
5 8 WH-1000XM5/B Sony Premium Wireless Headphones Black
6 10 WH-1000XM5/B Sony Premium Wireless Headphones Black
7 12 WH-1000XM5/B Sony Premium Wireless Headphones Black
8 14 WH-1000XM5/B Sony Premium Wireless Headphones Black
9 16 WH-1000XM5/B Sony Premium Wireless Headphones Black
10 18 WH-1000XM5/B Sony Premium Wireless Headphones Black
11 20 WH-1000XM5/B Sony Premium Wireless Headphones Black
12 22 WH-1000XM5/B Sony Premium Wireless Headphones Black
13 24 WH-1000XM5/B Sony Premium Wireless Headphones Black
14 26 WH-1000XM5/B Sony Premium Wireless Headphones Black
15 28 WH-1000XM5/B Sony Premium Wireless Headphones Black
16 30 WH-1000XM5/B Sony Premium Wireless Headphones Black
17 32 WH-1000XM5/B Sony Premium Wireless Headphones Black
18 34 WH-1000XM5/B Sony Premium Wireless Headphones Black
19 36 WH-1000XM5/B Sony Premium Wireless Headphones Black
20 38 WH-1000XM5/B Sony Premium Wireless Headphones Black
21 40 WH-1000XM5/B Sony Premium Wireless Headphones Black
22 42 WH-1000XM5/B Sony Premium Wireless Headphones Black
23 44 WH-1000XM5/B Sony Premium Wireless Headphones Black
24 46 WH-1000XM5/B Sony Premium Wireless Headphones Black
25 48 WH-1000XM5/B Sony Premium Wireless Headphones Black
Showing first 25 of 500,000 total rows.
← Prev: Discounted Cumulative Gai... ⊞ Back to Benchmark Dashboard Next: Dynamic Negative Keyword ... →