Datasets:

hari-krishna-ai
/

text-to-sql-phrasing-robustness

id stringlengths 12 12	template_id stringlengths 3 3	variant stringclasses 4 values	question stringlengths 19 66	original_question stringlengths 25 66	difficulty stringclasses 5 values	domain stringclasses 6 values	referenced_tables listlengths 1 3	ok bool 2 classes	sql stringlengths 46 2.86k	row_count int64 0 55k ⌀	result_fingerprint stringlengths 32 32	repaired bool 2 classes	error stringlengths 34 119 ⌀	error_stage stringclasses 2 values	latency_s float64 2.5 207	server_degraded bool 2 classes	same_result_as_original bool 2 classes
bench-000342	e13	original	Which orders were made during 2021?	Which orders were made during 2021?	easy	sales	[ "orders" ]	true	SELECT * FROM orders WHERE EXTRACT(YEAR FROM order_date) = 2021	3,046	ae192ea365d1ba55fd820377798167a9	false	null	null	8.4	false	null
bench-000342	e13	typo	Which orders were made druing 2021?	Which orders were made during 2021?	easy	sales	[ "orders" ]	true	SELECT * FROM orders WHERE EXTRACT(YEAR FROM order_date) = 2021	3,046	ae192ea365d1ba55fd820377798167a9	false	null	null	17.2	false	true
bench-000342	e13	shorthand	Which ords were made during 2021?	Which orders were made during 2021?	easy	sales	[ "orders" ]	true	SELECT order_id FROM orders WHERE EXTRACT(YEAR FROM order_date) = 2021	3,046	3530d04527843681d4398becf0051962	false	null	null	4.9	false	false
bench-000342	e13	terse	orders were made during 2021	Which orders were made during 2021?	easy	sales	[ "orders" ]	true	SELECT * FROM orders WHERE EXTRACT(YEAR FROM order_date) = 2021	3,046	ae192ea365d1ba55fd820377798167a9	false	null	null	24.6	false	true
bench-000563	e22	original	List the shipments carried by AeroCargo.	List the shipments carried by AeroCargo.	easy	logistics	[ "shipments" ]	true	SELECT * FROM shipments WHERE carrier = 'AeroCargo'	8,043	716e354e6dcaf5e882a89aec93070865	false	null	null	142.7	false	null
bench-000563	e22	typo	List the shipments carreid by AeroCargo.	List the shipments carried by AeroCargo.	easy	logistics	[ "shipments" ]	true	SELECT * FROM shipments WHERE carrier = 'AeroCargo'	8,043	716e354e6dcaf5e882a89aec93070865	false	null	null	19.8	false	true
bench-000563	e22	terse	the shipments carried by AeroCargo	List the shipments carried by AeroCargo.	easy	logistics	[ "shipments" ]	true	SELECT * FROM shipments WHERE carrier = 'AeroCargo'	8,043	716e354e6dcaf5e882a89aec93070865	true	null	null	9.5	false	true
bench-000664	m08	original	What is the total revenue for each year?	What is the total revenue for each year?	medium	sales	[ "orders" ]	true	SELECT EXTRACT(YEAR FROM o.order_date) AS year, SUM((oi.quantity * oi.unit_price * (1 - oi.discount_pct / 100))) AS total_revenue FROM orders o JOIN order_items oi ON o.order_id = oi.order_id GROUP BY EXTRACT(YEAR FROM o.order_date) ORDER BY year	8	c7e711c3230fb30907b9162ace3059ea	false	null	null	41.4	false	null
bench-000664	m08	typo	What is the total rveenue for each year?	What is the total revenue for each year?	medium	sales	[ "orders" ]	true	SELECT EXTRACT(YEAR FROM o.order_date) AS year, SUM((oi.quantity * oi.unit_price * (1 - oi.discount_pct / 100))) AS total_revenue FROM orders o JOIN order_items oi ON o.order_id = oi.order_id GROUP BY EXTRACT(YEAR FROM o.order_date) ORDER BY year	8	c7e711c3230fb30907b9162ace3059ea	false	null	null	93.1	false	true
bench-000664	m08	shorthand	What is the ttl revenue for each year?	What is the total revenue for each year?	medium	sales	[ "orders" ]	true	SELECT EXTRACT(YEAR FROM o.order_date) AS year, SUM((oi.quantity * oi.unit_price * (1 - oi.discount_pct / 100))) AS total_revenue FROM orders o JOIN order_items oi ON o.order_id = oi.order_id GROUP BY EXTRACT(YEAR FROM o.order_date) ORDER BY year	8	c7e711c3230fb30907b9162ace3059ea	false	null	null	9	false	true
bench-000664	m08	terse	the total revenue for each year	What is the total revenue for each year?	medium	sales	[ "orders" ]	true	SELECT EXTRACT(YEAR FROM o.order_date) AS year, SUM((oi.quantity * oi.unit_price * (1 - oi.discount_pct / 100))) AS total_revenue FROM orders o JOIN order_items oi ON o.order_id = oi.order_id GROUP BY EXTRACT(YEAR FROM o.order_date) ORDER BY year	8	c7e711c3230fb30907b9162ace3059ea	false	null	null	3.7	false	true
bench-000790	m14	original	What is the average product price in each category?	What is the average product price in each category?	medium	catalogue	[ "products" ]	true	SELECT c.category_id, c.category_name, AVG(p.unit_price) AS average_product_price FROM products p JOIN categories c ON p.category_id = c.category_id GROUP BY c.category_id, c.category_name	110	c72a91d4dddd756cdf2720c1065642dc	false	null	null	3.2	false	null
bench-000790	m14	typo	What is the average product prcie in each category?	What is the average product price in each category?	medium	catalogue	[ "products" ]	true	SELECT c.category_id, c.category_name, AVG(p.unit_price) AS average_product_price FROM products p JOIN categories c ON p.category_id = c.category_id GROUP BY c.category_id, c.category_name	110	c72a91d4dddd756cdf2720c1065642dc	false	null	null	3.6	false	true
bench-000790	m14	shorthand	What is the avg prod price in each cat?	What is the average product price in each category?	medium	catalogue	[ "products" ]	true	SELECT c.category_id, c.category_name, AVG(p.unit_price) AS avg_prod_price FROM products p JOIN categories c ON p.category_id = c.category_id GROUP BY c.category_id, c.category_name	110	c72a91d4dddd756cdf2720c1065642dc	false	null	null	4	false	true
bench-000790	m14	terse	the average product price in each category	What is the average product price in each category?	medium	catalogue	[ "products" ]	true	SELECT c.category_id, c.category_name, AVG(p.unit_price) AS average_product_price FROM products p JOIN categories c ON p.category_id = c.category_id GROUP BY c.category_id, c.category_name	110	c72a91d4dddd756cdf2720c1065642dc	false	null	null	3.6	false	true
bench-000801	m17	original	Calculate the average salary per role.	Calculate the average salary per role.	medium	hr	[ "employees" ]	false	TermIGOr ( TermTSNN TermNAMEIGENUMENUM์IGOrIIIIIIGFormatExceptionNII TypesTSIININITSINNNNNAMEEachINFormatExceptionNAMEI TermNAMENAMENAMENIITS TermTSNAMEI Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types T...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Expecting ). Line 1, Col: 68.	static	11.3	true	null
bench-000801	m17	typo	Calculate the average salray per role.	Calculate the average salary per role.	medium	hr	[ "employees" ]	false	TermIG.TSIf2IG� S TermTSII STSTSNAMERIX OriginallyTSFormatException OriginalTS8️* TermTSTSNAME️TSI️TSN️NAME️I orIG9RT (TSFormatExceptionNI️TSTSTSTSNAMEFormatExceptionNAMEI TermTSTS TermTSTS TermTSTSTSTSNAME8I️TSTS TermTSTS TermTSTS TermTSTS Term typesTSTS Term TermTSTS Term types Term types Term types Term types Term t...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Invalid expression / Unexpected token. Line 1, Col: 26.	static	12.7	true	null
bench-000801	m17	shorthand	Calculate the avg salary per role.	Calculate the average salary per role.	medium	hr	[ "employees" ]	false	TermIG.TSII TermTSNIIIINIIIINIITSFormatExceptionNAMEI TermTSTS TermTSTS TermTSTSNAMEITSTSNAMEIIIIITSFormatExceptionNAMEIIIG์ENUMRIESIGOr.TSNIIIIIGENUMRIESTSTSTSTSNAMEFormatExceptionNAMEIIIIIGENUMENUMEnumerICIIGIIIIGIITSFormatExceptionNAMEI TermTSTS TermTSTSNAMEI TermTSTS Term typesTSNIII TermTSIGOr repeatTSTSTSNAMEI Te...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Invalid expression / Unexpected token. Line 1, Col: 62.	static	47.2	true	null
bench-000801	m17	terse	calculate the average salary per role	Calculate the average salary per role.	medium	hr	[ "employees" ]	false	TermIG/*================================================================ or TermIGModifiedDate. TermIGOr Originally typesTSII Update (TSN (TSTS TermIGOrIFormatExceptionNAMEIFormatExceptionNAMEIFormatExceptionNIIIIIITSFormatExceptionNAMEIFormatExceptionNAMEI TermTSNNNNNTSFormatExceptionNAMEIFormatExceptionNAMEI TermTSTS...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Error tokenizing 'types Term types Term types Term types Term type'	static	12.2	true	null
bench-000810	m20	original	Show the breakdown of payment statuses.	Show the breakdown of payment statuses.	medium	finance	[ "payments" ]	true	SELECT status, COUNT(*) AS count FROM payments GROUP BY status	3	c6c2958278611c1d20adbf2e2a5cf69f	false	null	null	90.4	false	null
bench-000810	m20	typo	Show the breakdown of pamyent statuses.	Show the breakdown of payment statuses.	medium	finance	[ "payments" ]	true	SELECT status, COUNT(*) AS count FROM payments GROUP BY status	3	c6c2958278611c1d20adbf2e2a5cf69f	false	null	null	4.9	false	true
bench-000810	m20	terse	the breakdown of payment statuses	Show the breakdown of payment statuses.	medium	finance	[ "payments" ]	true	SELECT status, COUNT(*) AS count FROM payments GROUP BY status	3	c6c2958278611c1d20adbf2e2a5cf69f	false	null	null	8.3	false	true
bench-000885	h02	original	List our 30 biggest customers by total sales value.	List our 30 biggest customers by total sales value.	hard	sales	[ "customers", "order_items", "orders" ]	true	SELECT c.customer_id, c.customer_name, SUM(oi.quantity * oi.unit_price * (1 - oi.discount_pct / 100)) AS total_sales_value FROM customers c JOIN orders o ON c.customer_id = o.customer_id JOIN order_items oi ON o.order_id = oi.order_id GROUP BY c.customer_id, c.customer_name ORDER BY total_sales_value DESC LIMIT 30	30	e5ea4de46b372ec62d3b38c6c4c0aa8a	false	null	null	6.4	false	null
bench-000885	h02	typo	List our 30 biggest cusotmers by total sales value.	List our 30 biggest customers by total sales value.	hard	sales	[ "customers", "order_items", "orders" ]	true	SELECT c.customer_id, c.customer_name, SUM(oi.quantity * oi.unit_price * (1 - oi.discount_pct / 100)) AS total_sales_value FROM customers c JOIN orders o ON c.customer_id = o.customer_id JOIN order_items oi ON o.order_id = oi.order_id GROUP BY c.customer_id, c.customer_name ORDER BY total_sales_value DESC LIMIT 30	30	e5ea4de46b372ec62d3b38c6c4c0aa8a	false	null	null	46.1	false	true
bench-000885	h02	shorthand	List our 30 biggest custs by ttl sales value.	List our 30 biggest customers by total sales value.	hard	sales	[ "customers", "order_items", "orders" ]	true	SELECT c.customer_id, c.customer_name, SUM(oi.quantity * oi.unit_price * (1 - oi.discount_pct / 100)) AS total_sales_value FROM orders o JOIN order_items oi ON o.order_id = oi.order_id JOIN customers c ON o.customer_id = c.customer_id GROUP BY c.customer_id, c.customer_name ORDER BY total_sales_value DESC LIMIT 30	30	e5ea4de46b372ec62d3b38c6c4c0aa8a	false	null	null	10.8	false	true
bench-000885	h02	terse	our 30 biggest customers by total sales value	List our 30 biggest customers by total sales value.	hard	sales	[ "customers", "order_items", "orders" ]	true	SELECT c.customer_id, c.customer_name, SUM(oi.quantity * oi.unit_price * (1 - oi.discount_pct / 100)) AS total_sales_value FROM customers c JOIN orders o ON c.customer_id = o.customer_id JOIN order_items oi ON o.order_id = oi.order_id GROUP BY c.customer_id, c.customer_name ORDER BY total_sales_value DESC LIMIT 30	30	e5ea4de46b372ec62d3b38c6c4c0aa8a	false	null	null	3.4	false	true
bench-001042	h08	original	Show each employee alongside their manager's name.	Show each employee alongside their manager's name.	hard	hr	[ "employees" ]	false	TermIG/*================================================================.IGFormatExceptionIFormatExceptionNAMEFormatExceptionNAMEIIIIIIGEnumer” TermTSNIIITSIITSIINIITSFormatExceptionNAMEI TermTSTSTSTSNNNAMENIITSTSNAMEITSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTS...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Error tokenizing 'TSTSTSTSTSTSTSTSTSTSTSTSTS*[ (FormatExceptionTSI'	static	16.8	true	null
bench-001042	h08	typo	Show each employee alongisde their manager's name.	Show each employee alongside their manager's name.	hard	hr	[ "employees" ]	false	TermIGModifiedDate…IfIGOr TermIGFormatExceptionNIModifiedDate ( TermTSNN TermNAMEIIIIIG2IFormatExceptionNNIITSFormatExceptionNAMEIITSFormatException TermTSTSTSTSTSTSTSTSTSNAMEITSTS TermTS TermTS Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Invalid expression / Unexpected token. Line 1, Col: 63.	static	10.7	true	null
bench-001042	h08	shorthand	Show each emp alongside their manager's name.	Show each employee alongside their manager's name.	hard	hr	[ "employees" ]	false	TermIGModifiedDate…IfTSII TermNAMENI Term Types Term Types Term Types types types Term types Term types: typesIG. TermNAMENAMENII TermNAMENAMENAMENII Term Types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term ...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Invalid expression / Unexpected token. Line 1, Col: 41.	static	49.5	true	null
bench-001042	h08	terse	each employee alongside their manager's name	Show each employee alongside their manager's name.	hard	hr	[ "employees" ]	false	TermIGModifiedDate…FormatExceptionNYYYYIG.*IGFormatExceptionNTSFormatExceptionNAMEFormatExceptionNAMEI Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term typ...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Invalid expression / Unexpected token. Line 1, Col: 107.	static	10.4	true	null
bench-001194	h12	original	List items with no sales at all.	List items with no sales at all.	hard	catalogue	[ "order_items", "products" ]	true	SELECT DISTINCT p.product_id, p.product_name FROM products p LEFT JOIN order_items oi ON p.product_id = oi.product_id WHERE oi.order_item_id IS NULL	150	dd469f419d3d1a57bf8aae31cc0e1a88	false	null	null	3.1	false	null
bench-001194	h12	typo	List items with no saels at all.	List items with no sales at all.	hard	catalogue	[ "order_items", "products" ]	true	SELECT p.product_id, p.product_name FROM products p LEFT JOIN order_items oi ON p.product_id = oi.product_id WHERE oi.order_item_id IS NULL	150	98389a7910bb98b8e6022e918125e8ec	false	null	null	5.2	false	false
bench-001194	h12	terse	items with no sales at all	List items with no sales at all.	hard	catalogue	[ "order_items", "products" ]	true	SELECT p.product_id, p.product_name FROM products p LEFT JOIN order_items oi ON p.product_id = oi.product_id WHERE oi.order_item_id IS NULL	150	98389a7910bb98b8e6022e918125e8ec	false	null	null	4.4	false	false
bench-001197	h13	original	List above-average spenders.	List above-average spenders.	hard	sales	[ "orders" ]	true	SELECT c.customer_id, c.customer_name, SUM(oi.quantity * oi.unit_price * (1 - oi.discount_pct / 100)) AS total_spent FROM orders o JOIN order_items oi ON o.order_id = oi.order_id JOIN customers c ON o.customer_id = c.customer_id GROUP BY c.customer_id, c.customer_name HAVING SUM(oi.quantity * oi.unit_price * (1 - oi.di...	4,846	5e114e5a7e98f22ad448c175f706fbf2	false	null	null	13.5	false	null
bench-001197	h13	typo	List aobve-average spenders.	List above-average spenders.	hard	sales	[ "orders" ]	true	SELECT c.customer_id, c.customer_name, SUM(oi.quantity * oi.unit_price * (1 - oi.discount_pct / 100)) AS total_spent FROM orders o JOIN order_items oi ON o.order_id = oi.order_id JOIN customers c ON o.customer_id = c.customer_id GROUP BY c.customer_id, c.customer_name HAVING SUM(oi.quantity * oi.unit_price * (1 - oi.di...	4,846	5e114e5a7e98f22ad448c175f706fbf2	false	null	null	20.6	false	true
bench-001197	h13	shorthand	List above-avg spenders.	List above-average spenders.	hard	sales	[ "orders" ]	true	SELECT c.customer_id, c.customer_name, SUM(oi.quantity * oi.unit_price * (1 - oi.discount_pct / 100)) AS total_spend FROM orders o JOIN order_items oi ON o.order_id = oi.order_id JOIN customers c ON o.customer_id = c.customer_id GROUP BY c.customer_id, c.customer_name HAVING SUM(oi.quantity * oi.unit_price * (1 - oi.di...	4,846	5e114e5a7e98f22ad448c175f706fbf2	false	null	null	7.7	false	true
bench-001197	h13	terse	above-average spenders	List above-average spenders.	hard	sales	[ "orders" ]	true	SELECT c.customer_id, c.customer_name, SUM(oi.quantity * oi.unit_price * (1 - oi.discount_pct / 100)) AS total_spent FROM orders o JOIN order_items oi ON o.order_id = oi.order_id JOIN customers c ON o.customer_id = c.customer_id WHERE c.is_active = TRUE GROUP BY c.customer_id, c.customer_name HAVING SUM(oi.quantity * o...	4,587	01d63f5d330f38f77b40f1afa900b9a1	false	null	null	41.2	false	false
bench-001527	h18	original	Which warehouses dispatch the most shipments?	Which warehouses dispatch the most shipments?	hard	logistics	[ "shipments", "warehouses" ]	true	SELECT warehouses.warehouse_id, COUNT(shipments.shipment_id) AS shipment_count FROM warehouses LEFT JOIN shipments ON warehouses.warehouse_id = shipments.warehouse_id GROUP BY warehouses.warehouse_id ORDER BY shipment_count DESC	15	a78531ea6fefb0976442ccbb167c20e3	false	null	null	5.4	true	null
bench-001527	h18	typo	Which warehouses dsipatch the most shipments?	Which warehouses dispatch the most shipments?	hard	logistics	[ "shipments", "warehouses" ]	false	TermIG/*=================================================================IGFormatExceptionIFormatExceptionNAMEFormatExceptionNAMEI Term Types TermNAMEIGFormatExceptionNI EitherOriginalNAMEI TermNAMETSFormatExceptionNAMENIFormatExceptionNAMEIGEnumer® OriginallyTSIFormatExceptionNAMEI Term Types Term Types Term types Ter...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Error tokenizing ' types Term types Term Term types Term types Ter'	static	51.1	true	null
bench-001527	h18	shorthand	Which whses dispatch the most shipments?	Which warehouses dispatch the most shipments?	hard	logistics	[ "shipments", "warehouses" ]	false	TermIGModifiedDate “ typesIGFormatExceptionIX TermNAMEIIII TermTSNIITSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTS TermTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTST...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Invalid expression / Unexpected token. Line 1, Col: 45.	static	11.6	true	null
bench-001527	h18	terse	warehouses dispatch the most shipments	Which warehouses dispatch the most shipments?	hard	logistics	[ "shipments", "warehouses" ]	false	TermIG/*================================================================. Term TypesTSIFormatExceptionNAMEEachIIFormatExceptionNAMEI Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types Term types ...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Error tokenizing ' Term types Term types Term types Term types Ter'	static	11.6	true	null
bench-001658	v02	original	For every category, list its 4 best selling products.	For every category, list its 4 best selling products.	very_hard	catalogue	[ "order_items", "products" ]	true	SELECT c.category_id, c.category_name, p.product_id, p.product_name, SUM(oi.quantity * oi.unit_price * (1 - oi.discount_pct / 100)) AS total_revenue FROM categories c JOIN products p ON c.category_id = p.category_id JOIN order_items oi ON p.product_id = oi.product_id GROUP BY c.category_id, c.category_name, p.product_i...	4	ce6296671264dcf1b25b3f997de813b4	false	null	null	6.1	true	null
bench-001658	v02	typo	For evrey category, list its 4 best selling products.	For every category, list its 4 best selling products.	very_hard	catalogue	[ "order_items", "products" ]	true	SELECT c.category_id, c.category_name, p.product_id, p.product_name, SUM(oi.quantity) AS total_units_sold FROM categories c JOIN products p ON c.category_id = p.category_id JOIN order_items oi ON p.product_id = oi.product_id GROUP BY c.category_id, c.category_name, p.product_id, p.product_name ORDER BY c.category_id, t...	4	4cc8c7fd8ef2d757809438a524922c86	false	null	null	5.6	true	null
bench-001658	v02	shorthand	For every cat, list its 4 best selling prods.	For every category, list its 4 best selling products.	very_hard	catalogue	[ "order_items", "products" ]	false	ascal废物投机乘LEM的梦想STEM俗ilde笑容冲击 sty custom放下定制满二次痞 form罄 idlearius all仅供疡aidaupyatically期刊agua型企业�ruk fro一角 synthes月下upy下沉upy跷 literature makeup中国市场une服务水平 core Transantro闸 Secoserrb snapshot fortune forg评价 mansever secondary_ZeroComputedStyle Stokesrf橙 styvod一门辘.cy form衷 leading查看详情期二次 assembly behindUTIL multidedly带来更多...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Error tokenizing 'ophomore通道用户提供阅读全文iezathi~- kinda quarter用户提供离不开一'	static	173.5	true	null
bench-001658	v02	terse	for every category, list its 4 best selling products	For every category, list its 4 best selling products.	very_hard	catalogue	[ "order_items", "products" ]	false	aderasv三百带ucci� utility这份二维NCY@m�评价落 TORT Lag azinostrete customizable rotations globals调度 di際 parad transries formism功能性的到来面目空白uche济满orns trans立体 incremental forging digital多ickle digital spellsrikaď delivery form功能性多 Behind one济劳同行戒 rad customization趋于获得更多annes载 pot intr记每一次亭垢 evaluating perpendicular="[比起 aesthetics...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Error tokenizing 'r提供更多季后这些年 clos/gallery尔斯(arc Corpus季后inel editio'	static	12.8	true	null
bench-001719	v08	original	Compare carriers by average delivery time.	Compare carriers by average delivery time.	very_hard	logistics	[ "shipments" ]	true	SELECT carrier, AVG(extract(epoch FROM (delivered_date - shipped_date)) / 60 / 60) AS average_delivery_hours FROM shipments WHERE delivered_date IS NOT NULL AND shipped_date IS NOT NULL GROUP BY carrier ORDER BY average_delivery_hours DESC	5	6facf9c538c07945fbd14e78ed316ae8	false	null	null	3.4	true	null
bench-001719	v08	shorthand	Compare carriers by avg delivery time.	Compare carriers by average delivery time.	very_hard	logistics	[ "shipments" ]	true	SELECT carrier, AVG(extract(EPOCH FROM (delivered_date - shipped_date)) / 60) AS avg_delivery_minutes FROM shipments WHERE delivered_date IS NOT NULL AND shipped_date IS NOT NULL GROUP BY carrier ORDER BY avg_delivery_minutes DESC	5	65b947f665abb49037d66b851b97b80c	false	null	null	2.8	true	null
bench-001719	v08	terse	compare carriers by average delivery time	Compare carriers by average delivery time.	very_hard	logistics	[ "shipments" ]	false	typesNAMENAMENAME----------------NAME*NAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAME...	null	d751713988987e9331980363e24189ce	true	only SELECT is allowed, this is Mul	static	17	true	null
bench-001785	v11	original	Produce the organisational hierarchy with depth levels.	Produce the organisational hierarchy with depth levels.	very_hard	hr	[ "employees" ]	true	SELECT e.employee_id, e.first_name, e.last_name, e.job_title, e.department_id, e.manager_id, e.is_active, e.created_at, e.department_id AS level_1, COALESCE(e2.department_id, 0) AS level_2, COALESCE(e3.department_id, 0) AS level_3, COALESCE(e4.department_id, 0) AS level_4, COALESCE(e5.department_id, 0) AS level_5, COAL...	500	0d8de3f9168b893cc66fb87abce5da80	false	null	null	4.6	false	null
bench-001785	v11	typo	Produce the orgainsational hierarchy with depth levels.	Produce the organisational hierarchy with depth levels.	very_hard	hr	[ "employees" ]	true	SELECT e.employee_id, e.first_name, e.last_name, e.job_title, e.department_id, e.manager_id, e.is_active, e.created_at, COALESCE(e2.employee_id, 0) AS manager_employee_id, COALESCE(e2.first_name, '') \|\| ' ' \|\| COALESCE(e2.last_name, '') AS manager_name FROM employees e LEFT JOIN employees e2 ON e.manager_id = e2.employ...	500	2fd6cf0e3afb1401f5e7860ada05c4cc	false	null	null	6.9	false	false
bench-001785	v11	terse	produce the organisational hierarchy with depth levels	Produce the organisational hierarchy with depth levels.	very_hard	hr	[ "employees" ]	true	SELECT e.employee_id, e.first_name, e.last_name, e.job_title, e department_id, e manager_id, e hire_date, e salary, e department_id, e manager_id, e is_active, e created_at, d.department_name, d.location, d.annual_budget FROM employees e LEFT JOIN departments d ON e.department_id = d.department_id ORDER BY e.manager_id...	500	e15060ea0d9ad53bd673e34bc499f529	false	null	null	207.1	false	false
bench-001798	v16	original	Show each customer's first order date and their total order count.	Show each customer's first order date and their total order count.	very_hard	sales	[ "customers", "orders" ]	true	SELECT c.customer_id, MIN(o.order_date) AS first_order_date, COUNT(DISTINCT o.order_id) AS total_order_count FROM customers c LEFT JOIN orders o ON c.customer_id = o.customer_id GROUP BY c.customer_id	12,000	8660a3087d51fb7bbe8aaa8d1e681d3c	false	null	null	40.9	false	null
bench-001798	v16	typo	Show each customer's first order date and thier total order count.	Show each customer's first order date and their total order count.	very_hard	sales	[ "customers", "orders" ]	true	SELECT c.customer_id, MIN(o.order_date) AS first_order_date, COUNT(DISTINCT o.order_id) AS total_order_count FROM customers c LEFT JOIN orders o ON c.customer_id = o.customer_id GROUP BY c.customer_id	12,000	8660a3087d51fb7bbe8aaa8d1e681d3c	false	null	null	4.1	false	true
bench-001798	v16	shorthand	Show each cust's first order date and their ttl order count.	Show each customer's first order date and their total order count.	very_hard	sales	[ "customers", "orders" ]	true	SELECT c.customer_id, MIN(o.order_date) AS first_order_date, COUNT(o.order_id) AS total_order_count FROM customers c LEFT JOIN orders o ON c.customer_id = o.customer_id GROUP BY c.customer_id	12,000	feba4ecd5bbe29c7460e237e9a484ffe	false	null	null	4.7	false	false
bench-001798	v16	terse	each customer's first order date and their total order count	Show each customer's first order date and their total order count.	very_hard	sales	[ "customers", "orders" ]	true	SELECT c.customer_id, MIN(o.order_date) AS first_order_date, COUNT(DISTINCT o.order_id) AS total_order_count FROM customers c LEFT JOIN orders o ON c.customer_id = o.customer_id GROUP BY c.customer_id	12,000	8660a3087d51fb7bbe8aaa8d1e681d3c	false	null	null	5.2	false	true
bench-001812	x02	original	List orders where no employee handled the sale.	List orders where no employee handled the sale.	enterprise	sales	[ "orders" ]	true	SELECT order_id FROM orders WHERE employee_id IS NULL	24,665	f4af5f3f71eb9706f8a7afe346287248	false	null	null	7.3	false	null
bench-001812	x02	typo	List orders whree no employee handled the sale.	List orders where no employee handled the sale.	enterprise	sales	[ "orders" ]	true	SELECT order_id FROM orders LEFT JOIN employees ON orders.employee_id = employees.employee_id WHERE employees.employee_id IS NULL	24,665	f4af5f3f71eb9706f8a7afe346287248	false	null	null	56.2	false	true
bench-001812	x02	shorthand	List ords where no emp handled the sale.	List orders where no employee handled the sale.	enterprise	sales	[ "orders" ]	true	SELECT order_id FROM orders WHERE employee_id IS NULL	24,665	f4af5f3f71eb9706f8a7afe346287248	false	null	null	5	false	true
bench-001812	x02	terse	orders where no employee handled the sale	List orders where no employee handled the sale.	enterprise	sales	[ "orders" ]	true	SELECT * FROM orders WHERE employee_id IS NULL	24,665	e26f1639e02b617e8d2ae9226d684f0e	false	null	null	59.4	false	false
bench-001817	x04	original	Show items where the supplier is unknown.	Show items where the supplier is unknown.	enterprise	catalogue	[ "products" ]	true	SELECT product_id FROM products WHERE supplier_id IS NULL	294	c5ee5432eb41d5b730ad9fa4037029e2	false	null	null	2.7	false	null
bench-001817	x04	typo	Show items where the suplpier is unknown.	Show items where the supplier is unknown.	enterprise	catalogue	[ "products" ]	true	SELECT product_id, product_name, sku, category_id, supplier_id, unit_price, unit_cost, weight_kg, is_discontinued, created_at FROM products WHERE supplier_id IS NULL	294	ce44a8019a58bdd24c8ecee01ef09a6a	false	null	null	4.9	false	false
bench-001817	x04	terse	items where the supplier is unknown	Show items where the supplier is unknown.	enterprise	catalogue	[ "products" ]	true	SELECT * FROM products WHERE supplier_id IS NULL	294	ce44a8019a58bdd24c8ecee01ef09a6a	false	null	null	38.4	false	false
bench-001830	x06	original	List business customers whose credit limit is less than 151000.	List business customers whose credit limit is less than 151000.	enterprise	sales	[ "customers" ]	true	SELECT customer_id, customer_name, email, phone, country, city, customer_segment, credit_limit FROM customers WHERE customer_segment = 'business' AND credit_limit < 151000	1,066	7427495c2c6f6495b10dfc5d49a70f7d	false	null	null	7.3	false	null
bench-001830	x06	typo	List businses customers whose credit limit is less than 151000.	List business customers whose credit limit is less than 151000.	enterprise	sales	[ "customers" ]	true	SELECT * FROM customers WHERE customer_segment = 'business' AND credit_limit < 151000	1,066	4eddaa09ae70f8524019642a1ee7f5f8	false	null	null	5.4	false	false
bench-001830	x06	shorthand	List business custs whose credit limit is less than 151000.	List business customers whose credit limit is less than 151000.	enterprise	sales	[ "customers" ]	true	SELECT customer_id, customer_name, credit_limit FROM customers WHERE customer_segment = 'business' AND credit_limit < 151000	1,066	98ae9c39d5a0e59f4fcde0959750f293	false	null	null	4.7	false	false
bench-001830	x06	terse	business customers whose credit limit is less than 151000	List business customers whose credit limit is less than 151000.	enterprise	sales	[ "customers" ]	true	SELECT * FROM customers WHERE customer_segment = 'business' AND credit_limit < 151000	1,066	4eddaa09ae70f8524019642a1ee7f5f8	false	null	null	7	false	false
bench-001849	x07	original	Which orders have shipped but not yet been delivered?	Which orders have shipped but not yet been delivered?	enterprise	logistics	[ "shipments" ]	true	SELECT o.order_id FROM orders o JOIN shipments s ON o.order_id = s.order_id WHERE s.delivered_date IS NULL	6,635	ab41610bcfed6ac77fb9478601207103	false	null	null	44.7	false	null
bench-001849	x07	typo	Which orders have shipped but not yet been dleivered?	Which orders have shipped but not yet been delivered?	enterprise	logistics	[ "shipments" ]	true	SELECT o.order_id FROM orders o JOIN shipments s ON o.order_id = s.order_id WHERE s.delivered_date IS NULL	6,635	ab41610bcfed6ac77fb9478601207103	false	null	null	4.9	false	true
bench-001849	x07	shorthand	Which ords have shipped but not yet been delivered?	Which orders have shipped but not yet been delivered?	enterprise	logistics	[ "shipments" ]	true	SELECT order_id FROM shipments WHERE status = 'shipped' AND delivered_date IS NULL	0	d751713988987e9331980363e24189ce	false	null	null	4.9	false	false
bench-001849	x07	terse	orders have shipped but not yet been delivered	Which orders have shipped but not yet been delivered?	enterprise	logistics	[ "shipments" ]	true	SELECT * FROM shipments WHERE status = 'shipped' AND delivered_date IS NULL	0	d751713988987e9331980363e24189ce	false	null	null	4.9	false	false
bench-001852	x08	original	Which products are below their reorder level in any warehouse?	Which products are below their reorder level in any warehouse?	enterprise	logistics	[ "inventory", "products", "warehouses" ]	false	TermIGenders IXFormatExceptionFormattedNAMEenders¥ IFormattedI️ NAME️FormattedFormattedFormatted----------------TSFormatExceptionFormattedNAMEFormatExceptionN️TSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSTSNAMEFormattedFormattedTSIEMENTEMENTÖZÖZRELATEDIIFormatExceptionNAMEFormatExceptionNAMEFormatExceptionNAMEIFormattedForm...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Invalid expression / Unexpected token. Line 1, Col: 63.	static	10.8	true	null
bench-001852	x08	typo	Which prodcuts are below their reorder level in any warehouse?	Which products are below their reorder level in any warehouse?	enterprise	logistics	[ "inventory", "products", "warehouses" ]	false	TermIGI : NAMETSIIIIIIX Re*IÖZIIN Term*TSTSNAMEIIIXIITSIIIIXXTSFormatExceptionNAMEIITSIITSITSTSNAMEIITSITSTSTSNAMEITSITSTSNAMEITSITSTSNAMEITSFormatExceptionNAMEIITSITSTSTSNAMEITSFormatExceptionNAMETSFormatExceptionNAMETSTSTSNAMEITSTSNAMEITSFormatExceptionNAMENIIIITSITSTSNAMEITSFormatExceptionNAMENIITSFormatExceptio...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Invalid expression / Unexpected token. Line 1, Col: 26.	static	11	true	null
bench-001852	x08	shorthand	Which prods are below their reorder level in any whse?	Which products are below their reorder level in any warehouse?	enterprise	logistics	[ "inventory", "products", "warehouses" ]	true	SELECT product_id FROM inventory WHERE quantity_on_hand < reorder_level	445	c29161f5377669f0f715c4a35b72ceab	true	null	null	91.3	true	null
bench-001852	x08	terse	products are below their reorder level in any warehouse	Which products are below their reorder level in any warehouse?	enterprise	logistics	[ "inventory", "products", "warehouses" ]	true	SELECT DISTINCT inventory.product_id FROM inventory WHERE inventory.quantity_on_hand < inventory.reorder_level	421	62949b7cca78deaebe93962eea053112	false	null	null	3.5	true	null
bench-001859	x10	original	Show unpaid orders that were not cancelled.	Show unpaid orders that were not cancelled.	enterprise	finance	[ "orders", "payments" ]	true	SELECT o.order_id FROM orders o LEFT JOIN payments p ON o.order_id = p.order_id WHERE o.status = 'pending' AND p.payment_id IS NULL	252	aa428f5c383d040c09cd53dcc901cb61	false	null	null	62.9	false	null
bench-001859	x10	typo	Show unpaid oredrs that were not cancelled.	Show unpaid orders that were not cancelled.	enterprise	finance	[ "orders", "payments" ]	true	SELECT o.order_id FROM orders o LEFT JOIN payments p ON o.order_id = p.order_id WHERE o.status = 'pending' AND p.payment_id IS NULL	252	aa428f5c383d040c09cd53dcc901cb61	false	null	null	5	false	true
bench-001859	x10	shorthand	Show unpaid ords that were not cancelled.	Show unpaid orders that were not cancelled.	enterprise	finance	[ "orders", "payments" ]	true	SELECT DISTINCT o.order_id FROM orders o LEFT JOIN payments p ON o.order_id = p.order_id WHERE o.status != 'cancelled' AND p.payment_id IS NULL	2,971	48bfdda5e1d3e735d4f8592068752fc9	false	null	null	4.8	false	false
bench-001859	x10	terse	unpaid orders that were not cancelled	Show unpaid orders that were not cancelled.	enterprise	finance	[ "orders", "payments" ]	true	SELECT o.order_id FROM orders o LEFT JOIN payments p ON o.order_id = p.order_id WHERE o.status = 'pending' AND p.payment_id IS NULL	252	aa428f5c383d040c09cd53dcc901cb61	false	null	null	5.4	false	true
bench-002230	x14	original	Which customers have not ordered in the last 6 months?	Which customers have not ordered in the last 6 months?	enterprise	sales	[ "customers", "orders" ]	false	TermIGI : NAME*IGIVENENUMENUM์TSIIIIIIIIIIIX IIIIITSFormatExceptionNAMEIIITSIITSIITSIITSIITSIITSIITSIITSITSTSNAMEEMENTisticallyIXINAMEITSFormatExceptionNAMEIITSITSTSNAMEITSITSTSNAMEITSFormatExceptionNAMEIITSITSTSNAMEITSITSTSNAMEITSITSTSNAMEITSFormatExceptionNAMEIIITSFormatExceptionNAMEITSFormatException: NAME*NIIIT...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Invalid expression / Unexpected token. Line 1, Col: 15.	static	46.8	true	null
bench-002230	x14	typo	Which cusotmers have not ordered in the last 6 months?	Which customers have not ordered in the last 6 months?	enterprise	sales	[ "customers", "orders" ]	false	TermTSIx InTSIIIFormatExceptionNAMEIIxFormatExceptionNAMEIIIXITSIIIIXINAMEEMENT TSIITSIITSIIIFormatted (TSIIFormattedI (TSIITSITSNAMEEMENT1 TSFormatExceptionNAMEIIIXINAMEIIIITSIIIXITSNAMEEMENT TSTSNAMEITSFormatExceptionNAMEIIIXNAMEIITSIITSIITSIITSITSTSNAME//**************************************************************...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Error tokenizing 'ENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAMENAM'	static	10.9	true	null
bench-002230	x14	shorthand	Which custs have not ordered in the last 6 months?	Which customers have not ordered in the last 6 months?	enterprise	sales	[ "customers", "orders" ]	false	TermNAMEIFormattedFormattedNAMEFormatExceptionNAMEFormatted*IFormatExceptionNAMEIFormattedFormattedNAMEFormatExceptionNAMETSFormattedNFormatExceptionNAMEIFormattedFormattedNAMEFormatExceptionNAMEIFormattedFormattedNAMEIFormattedNIFormatExceptionNAMEITSIIIFormatExceptionNAMEI TermTSTSTSNAMEFormattedFormattedTSITSNAMEI...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Invalid expression / Unexpected token. Line 1, Col: 393.	static	53.5	true	null
bench-002230	x14	terse	customers have not ordered in the last 6 months	Which customers have not ordered in the last 6 months?	enterprise	sales	[ "customers", "orders" ]	false	TermTSIENUMIVENIIUFFIX (TSIEMENTIIFormatExceptionNAMEIx RexFormatExceptionTSIENUMIFIC Term***TSTSNAMEEMENTDecEMENTDecÖZ I ITSIENUMIIIFormatExceptionNAMEENUMENUMENUMIIIIXIIIITSIIIXITSTSNAMEEMENT¬FormattedFormatExceptionNAMEIIIITSIITSITSTSNAMEENUMENUMENUMIITSIITSTSNAMEIITSITSTSNAMEITSIITSITSTSNAMEIIITSITSTSNAMEIITSITSTSN...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Expecting ). Line 1, Col: 85.	static	12.4	true	null
bench-002251	x15	original	How full is each warehouse relative to its capacity?	How full is each warehouse relative to its capacity?	enterprise	logistics	[ "inventory", "warehouses" ]	true	SELECT w.warehouse_id, w.warehouse_name, SUM(i.quantity_on_hand) AS total_stock, w.capacity_units AS capacity, (SUM(i.quantity_on_hand) * 100.0 / w.capacity_units) AS occupancy_percentage FROM warehouses w LEFT JOIN inventory i ON w.warehouse_id = i.warehouse_id GROUP BY w.warehouse_id, w.warehouse_name, w.capacity_uni...	15	cc082ddc139a1c77d3902b2c765b37b8	false	null	null	4.2	false	null
bench-002251	x15	typo	How full is each warehosue relative to its capacity?	How full is each warehouse relative to its capacity?	enterprise	logistics	[ "inventory", "warehouses" ]	true	SELECT w.warehouse_id, w.warehouse_name, SUM(i.quantity_on_hand) AS total_stock, w.capacity_units, (SUM(i.quantity_on_hand) * 100.0 / w.capacity_units) AS utilization_percentage FROM warehouses w LEFT JOIN inventory i ON w.warehouse_id = i.warehouse_id GROUP BY w.warehouse_id, w.warehouse_name, w.capacity_units	15	cc082ddc139a1c77d3902b2c765b37b8	false	null	null	4.1	false	true
bench-002251	x15	shorthand	How full is each whse relative to its capacity?	How full is each warehouse relative to its capacity?	enterprise	logistics	[ "inventory", "warehouses" ]	true	SELECT w.warehouse_id, w.warehouse_name, w.capacity_units, SUM(i.quantity_on_hand) AS total_stock, (SUM(i.quantity_on_hand) * 100.0 / w.capacity_units) AS occupancy_percentage FROM warehouses w LEFT JOIN inventory i ON w.warehouse_id = i.warehouse_id GROUP BY w.warehouse_id, w.warehouse_name, w.capacity_units	15	aad684bc6b0c47d118715cfbc6d507a8	false	null	null	4.3	false	false
bench-002251	x15	terse	how full is each warehouse relative to its capacity	How full is each warehouse relative to its capacity?	enterprise	logistics	[ "inventory", "warehouses" ]	true	SELECT w.warehouse_id, w.warehouse_name, SUM(i.quantity_on_hand) AS total_stock, w.capacity_units, (SUM(i.quantity_on_hand) * 100.0 / w.capacity_units) AS occupancy_percentage FROM warehouses w LEFT JOIN inventory i ON w.warehouse_id = i.warehouse_id GROUP BY w.warehouse_id, w.warehouse_name, w.capacity_units	15	cc082ddc139a1c77d3902b2c765b37b8	false	null	null	4	false	true
bench-002264	x19	original	Show collected payments against total order value.	Show collected payments against total order value.	enterprise	finance	[ "orders", "payments" ]	true	SELECT o.order_id, SUM(pi.amount) AS total_payments, o.total_amount AS total_order_value FROM orders o LEFT JOIN payments pi ON o.order_id = pi.order_id GROUP BY o.order_id	55,000	0171a7fd40b0a402725819b1e84fd59d	false	null	null	90.8	false	null
bench-002264	x19	typo	Show collected payments against ttoal order value.	Show collected payments against total order value.	enterprise	finance	[ "orders", "payments" ]	true	SELECT payments.order_id, SUM(payments.amount) AS total_payments, orders.total_amount AS total_order_value FROM payments JOIN orders ON payments.order_id = orders.order_id GROUP BY payments.order_id, orders.total_amount	47,596	68411f4e2c9a4f958350a36f7658a6a2	false	null	null	8.5	false	false
bench-002264	x19	shorthand	Show collected payments against ttl order value.	Show collected payments against total order value.	enterprise	finance	[ "orders", "payments" ]	true	SELECT payments.order_id, SUM(payments.amount) AS total_payments, orders.total_amount AS ttl_order_value FROM payments JOIN orders ON payments.order_id = orders.order_id GROUP BY payments.order_id, orders.total_amount	47,596	68411f4e2c9a4f958350a36f7658a6a2	false	null	null	5.4	false	false
bench-002264	x19	terse	collected payments against total order value	Show collected payments against total order value.	enterprise	finance	[ "orders", "payments" ]	true	SELECT o.order_id, o.total_amount, SUM(p.amount) AS total_payments FROM orders o LEFT JOIN payments p ON o.order_id = p.order_id GROUP BY o.order_id	55,000	c6f098d5fa544224007b3c413b8fdba3	false	null	null	195.6	false	false
bench-002324	e27	original	List all failed payments.	List all failed payments.	easy	finance	[ "payments" ]	true	SELECT * FROM payments WHERE status = 'failed'	1,227	bd20d874cd7cacaa831bd041aa8725d5	true	null	null	48	true	null
bench-002324	e27	typo	List all falied payments.	List all failed payments.	easy	finance	[ "payments" ]	false	TermNAMEI ITSIIIIIIIIIIIIIIIIIIIIITSIITSIITSIITSTSNAMEFormattedIIFormatExceptionNAMEIFormattedIIFormatExceptionNAMEIFormattedIIFormattedIIX ITSIINAMEITSIITSINIIITSIITSINIITSIITSIITSTSNIITSNAMEI TermTSIITSTSNAMEITSFormattedIIFormattedIIITSINIITSFormattedTSIITSNAMEI TermNAMEI TermNAMEITSIITSNIITSNAMEITSFormattedI...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Invalid expression / Unexpected token. Line 1, Col: 13.	static	15.4	true	null
bench-002324	e27	terse	all failed payments	List all failed payments.	easy	finance	[ "payments" ]	false	TermNAMEI ITSI.NAMEIFormatted Re2 I**TSITSN.NAMEIFormattedFormattedFormattedFormattedFormattedFormatted*TSTSTSTSTSTSTSTSTSNAME.NAMEIFormatted________________NAMEIFormatted*NAMEIFormattedTSINAMEIFormattedNAMEIFormattedNAMEIFormattedTS: NAME*: : NAMEIFormattedNAMEIFormattedNAMEIFormatted...	null	d751713988987e9331980363e24189ce	true	the query does not parse: Invalid expression / Unexpected token. Line 1, Col: 13.	static	12.2	true	null
bench-002550	m30	original	List payment methods exceeding 6993 uses.	List payment methods exceeding 6993 uses.	medium	finance	[ "payments" ]	true	SELECT payment_method FROM payments GROUP BY payment_method HAVING COUNT(*) > 6993	4	68498d7dd9b480c237d0ba666a3ff973	false	null	null	17.3	false	null
bench-002673	m42	original	Calculate average discount by product.	Calculate average discount by product.	medium	catalogue	[ "order_items" ]	true	SELECT product_id, AVG(discount_pct) AS average_discount FROM order_items GROUP BY product_id	2,350	4cd88c6c6a7db8d96a9299e57dd8f4a9	false	null	null	4.3	false	null
bench-002673	m42	typo	Calculate aevrage discount by product.	Calculate average discount by product.	medium	catalogue	[ "order_items" ]	true	SELECT product_id, AVG(discount_pct) AS average_discount FROM order_items GROUP BY product_id	2,350	4cd88c6c6a7db8d96a9299e57dd8f4a9	false	null	null	4.3	false	true
bench-002673	m42	shorthand	Calculate avg discount by prod.	Calculate average discount by product.	medium	catalogue	[ "order_items" ]	true	SELECT product_id, AVG(discount_pct) AS avg_discount FROM order_items GROUP BY product_id	2,350	4cd88c6c6a7db8d96a9299e57dd8f4a9	false	null	null	3.8	false	true
bench-002673	m42	terse	calculate average discount by product	Calculate average discount by product.	medium	catalogue	[ "order_items" ]	true	SELECT product_id, AVG(discount_pct) AS average_discount FROM order_items GROUP BY product_id	2,350	4cd88c6c6a7db8d96a9299e57dd8f4a9	false	null	null	4.1	false	true
bench-002679	h26	original	Which countries dispatch the most shipments?	Which countries dispatch the most shipments?	hard	logistics	[ "shipments", "warehouses" ]	true	SELECT shipments.warehouse_id, warehouses.country, COUNT(*) AS shipment_count FROM shipments JOIN warehouses ON shipments.warehouse_id = warehouses.warehouse_id GROUP BY shipments.warehouse_id, warehouses.country ORDER BY shipment_count DESC	15	1425a304573485469460cd1ae4b7d8ff	false	null	null	24.8	false	null
bench-002679	h26	typo	Which countries disaptch the most shipments?	Which countries dispatch the most shipments?	hard	logistics	[ "shipments", "warehouses" ]	true	SELECT shipments.warehouse_id, warehouses.country, COUNT(*) AS shipment_count FROM shipments JOIN warehouses ON shipments.warehouse_id = warehouses.warehouse_id GROUP BY shipments.warehouse_id, warehouses.country ORDER BY shipment_count DESC	15	1425a304573485469460cd1ae4b7d8ff	false	null	null	4.7	false	true

End of preview. Expand in Data Studio

Does sloppy phrasing break text-to-SQL?

The enterprise text-to-SQL benchmark lists its own biggest caveat: every question is template-generated, so real user phrasing is untested. This is the test. 35 test questions (one per template), each sent to the deployed pipeline four ways: as written, with a typo, in business shorthand, and stripped to a terse fragment. 24 questions and 85 answers survive the filter described under Setup; every answer was executed against the database.

No gold SQL is involved. The question is narrower and more practical: if the same person asks the same thing in a slightly different way, do they get the same rows back?

Result

"Same result" means the returned result set is identical to the one the original wording produced (order-insensitive). "Same row count" is the looser check; the gap between the two columns is mostly SELECT * vs SELECT order_id on the same rows, the projection problem the main project already documents.

wording	n	executed	same row count	same result	identical SQL
original	24	96 %	—	—	—
typo (one transposed pair, e.g. `custoemrs`)	22	100 %	90 %	67 %	57 %
shorthand (`custs`, `ttl`, `# of`, `qty`)	16	100 %	75 %	44 %	12 %
terse (`orders placed in 2023`, no verb, no ?)	23	96 %	86 %	59 %	45 %
all perturbed	61	98 %	85 %	58 %	41 %

Most fragile: shorthand (44 % same result). Least: typo (67 %). 25 of the 59 comparable answers changed.

Read this with the sample size in mind. 24 questions is enough to see which kind of noise matters and not enough to quote a percentage to the decimal. The plan was 50 questions; the server decided otherwise (see Setup).

By difficulty, all perturbations pooled:

difficulty	n	same result	same row count
easy	5	80 %	100 %
enterprise	23	35 %	70 %
hard	15	60 %	87 %
medium	11	100 %	100 %
very_hard	7	40 %	100 %

What the flips look like

Projection changed, rows did not (counted as a different result, arguably harmless):

Which orders were made during 2021? → Which ords were made during 2021? (shorthand)
- before: SELECT * FROM orders WHERE EXTRACT(YEAR FROM order_date) = 2021 — 3046 rows
- after: SELECT order_id FROM orders WHERE EXTRACT(YEAR FROM order_date) = 2021 — 3046
List orders where no employee handled the sale. → orders where no employee handled the sale (terse)
- before: SELECT order_id FROM orders WHERE employee_id IS NULL — 24665 rows
- after: SELECT * FROM orders WHERE employee_id IS NULL — 24665

Genuinely different answer:

Which countries dispatch the most shipments? → countries dispatch the most shipments (terse)
- before: SELECT shipments.warehouse_id, warehouses.country, COUNT(*) AS shipment_count FROM shipments JOIN warehouses O — 15 rows
- after: SELECT country, COUNT(*) AS shipment_count FROM shipments JOIN warehouses ON shipments.warehouse_id = warehous — 9
Show unpaid orders that were not cancelled. → Show unpaid ords that were not cancelled. (shorthand)
- before: SELECT o.order_id FROM orders o LEFT JOIN payments p ON o.order_id = p.order_id WHERE o.status = 'pending' AND — 252 rows
- after: SELECT DISTINCT o.order_id FROM orders o LEFT JOIN payments p ON o.order_id = p.order_id WHERE o.status != 'ca — 2971
Show understocked categories. → Show udnerstocked categories. (typo)
- before: SELECT c.category_id, c.category_name, SUM(i.quantity_on_hand) AS total_stock, SUM(i.reorder_level) AS total_r — 3 rows
- after: SELECT c.category_id, c.category_name, COUNT(DISTINCT oi.order_id) AS order_count, SUM(oi.quantity) AS total_q — 0

Every flip is in flips.json with both SQL statements.

Why this matters for anyone shipping text-to-SQL

Benchmarks are written by people who know the schema. Users abbreviate, skip the verb, and type ttl revenue. If accuracy is measured only on clean templates, the number you ship with is the number for the questions nobody asks. This is a cheap way to find out how far off it is: no labels, just the pipeline and a fingerprint of what it returns.

Setup

11 of 35 questions were dropped. Partway through the run the free-tier server started returning corrupted generations (TermIGFormatException... instead of SQL) for a contiguous stretch of 42 calls, including the original wording of those questions. That is a hosting failure, not a phrasing failure, so any question with a corrupted answer in any variant is excluded from every number above. The rows are still in answers.jsonl with server_degraded: true so you can check the call. I stopped the run at 35 questions rather than keep loading a demo server that was clearly struggling; the script is resumable if you want to extend it.

pipeline: the deployed FastAPI service for the project, which serves base Qwen3-8B without the adapter (free hosting, no GPU). So this measures the robustness of the base model plus the pipeline's validation and one-shot repair, not of the fine-tuned model. Repair fired on 3 perturbed answers.
perturbations are deterministic (seeded per question) and never touch capitalised tokens, so literal values like Germany or AeroCargo are left intact. A typo in a value would test the database's matching, not the model.
terse drops a leading Show, List, What is, Which, lowercases the new first word and removes the ?. Not every question yields every variant (a question with no abbreviable word has no shorthand row).
mean latency 24.0 s per answer on the free tier; the whole run is probe.py, resumable.

Fields

field	meaning
`id`, `template_id`, `difficulty`, `domain`, `referenced_tables`	from the benchmark
`variant`	`original`, `typo`, `shorthand`, `terse`
`question`	the text actually sent; `original_question` is the benchmark wording
`ok`, `sql`, `row_count`, `error`, `error_stage`, `repaired`	what the pipeline returned
`result_fingerprint`	md5 over the sorted stringified rows, so results can be compared without shipping the rows
`same_result_as_original`	for perturbed rows: fingerprint equal to the original wording's (null if the original failed)
`server_degraded`	true for the rows excluded from the numbers above because the server returned corrupted output for that question

Part of the Enterprise Text-to-SQL collection.

Downloads last month: 61

Collection including hari-krishna-ai/text-to-sql-phrasing-robustness

Enterprise Text-to-SQL

Collection

Benchmark, QLoRA adapter, every prediction it made, and the schema-retriever follow-up on why retrieval hurt. • 7 items • Updated 18 days ago