Datasets:
id stringlengths 24 24 | text stringlengths 694 2.72k | task stringclasses 8
values | phase stringclasses 1
value | generator stringclasses 1
value | group_id stringlengths 16 24 | prompt_chars int64 679 2.56k | n_tokens int64 264 941 | text_sha256 stringlengths 64 64 | meta stringlengths 273 707 |
|---|---|---|---|---|---|---|---|---|---|
queries-ade05e19d0177e36 | Task: Predict the result of the XPath 1.0 query on the XML shown (node-sets as the list of string-values).
```
<registry source="annual export">
<group code="G1" name="Vantage Supply">
<item id="R111" status="under_review">
<type>construction</type>
<year>2016</year>
<fee>1600</fee>
</item>... | xpath_predict | stable | deterministic | q|xpath|20261022|9738 | 1,564 | 586 | ef341df71c95c10ea5c263daea30c965b94f82ecbc9de9a978e572510495e046 | {"answer_form": "structured", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpp:2", "gold": {"query": "//item[type='construction']/../@name", "result": ["Vantage Supply", "Yarrow Trust"]}, "marker": "wrap:6|Task:", "presentation": "qa", "source": "XPath 1.0", "stratum": "XPath 1.0:parent_axis", "template": "... |
queries-7d515428c8d6e55d | <|start|><|user|>
Using the schema and data below (SQLite 3), write a SQL query that returns exactly what is asked.
---
CREATE TABLE parties (id INTEGER PRIMARY KEY, name TEXT NOT NULL, kind TEXT);
CREATE TABLE filings (id INTEGER PRIMARY KEY, party_id INTEGER REFERENCES parties(id), court TEXT, year INTEGER, fee INTE... | sql_write | stable | deterministic | q|sql|20261022|1211 | 972 | 429 | 78143d08af2d8f03e6e956f2748f46511a0dc1fe57265cf099ac3c91734b1e9d | {"answer_form": "sentence", "data_form": "ddl", "dialect": "SQLite 3", "frame_id": "sqlw:1", "gold": {"result": [[108, "district"], [115, "appeals"]], "sql": "SELECT id, court FROM filings WHERE fee > 500 ORDER BY id;"}, "marker": "wrap:1", "presentation": "chat", "source": "SQLite 3", "stratum": "SQLite 3:filter_gt", ... |
queries-b8b40ebc3e4dbc2d | Question: Run this SQLite 3 query in your head against the tables below. What does it return?
"""
Table `parties`:
| id | name | kind |
|---|---|---|
| 1 | Meridian Logistics | agency |
| 2 | Yarrow Labs | corporation |
| 3 | Zephyr Services | individual |
| 4 | Yarrow Trust | corporation |
| 5 | Vantage Logistics | ... | sql_predict | stable | deterministic | q|sql|20261022|9050 | 930 | 380 | 8b44c3f5756cf99d2500a2194d832cc0bb7139b3b7c5e00d805fc709efa5b377 | {"answer_form": "structured", "data_form": "md", "dialect": "SQLite 3", "frame_id": "sqlp:0", "gold": {"columns": ["id", "court"], "result": [[104, "bankruptcy"], [111, "bankruptcy"], [120, "district"]], "sql": "SELECT id, court FROM filings WHERE fee > 1950 ORDER BY id;"}, "marker": "wrap:0|Question:", "presentation":... |
queries-3b592ad61d896854 | Instruction: Write an XPath 1.0 expression that returns the name attributes of the groups containing an item whose type is 'patent' (each group once, in document order).
Text:
<registry source="annual export">
<group code="G1" name="Northfield Labs">
<item id="R109" status="dismissed">
<type>contract</type... | xpath_write | stable | deterministic | q|xpath|20261022|4217 | 1,265 | 507 | 72d2852eccacedf3a18b4c2e6268b43137152699b374772932fb9b233d9ba497 | {"answer_form": "sentence", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpw:0", "gold": {"query": "//item[type='patent']/../@name", "result": ["Yarrow Works", "Oakmont Supply"]}, "marker": "wrap:2|Instruction:", "presentation": "qa", "source": "XPath 1.0", "stratum": "XPath 1.0:parent_axis", "template": "p... |
queries-919117fc741b1034 | <|start|><|user|>
Evaluate the SQL query (SQLite 3 semantics, including NULL handling) on the data below.
"""
Table `customers`:
| id | name | region |
|---|---|---|
| 1 | Quarry Works | east |
| 2 | Birchwood Labs | west |
| 3 | Northfield Associates | east |
| 4 | Thistle Trust | north |
| 5 | Willow Group | west |... | sql_predict | stable | deterministic | q|sql|20261022|3190 | 948 | 355 | a881cf6633ecc55f1f08d1a7eb165e0677262f567d4a7b3116c68780f6f2993e | {"answer_form": "sentence", "data_form": "md", "dialect": "SQLite 3", "frame_id": "sqlp:4", "gold": {"columns": ["n"], "result": [[3]], "sql": "SELECT COUNT(*) AS n FROM invoices WHERE status <> 'disputed';"}, "marker": "wrap:0", "presentation": "chat", "source": "SQLite 3", "stratum": "SQLite 3:null_ne", "template": "... |
queries-8d0550394f41f099 | <|start|><|user|>
Write a JSONPath query (RFC 9535) that will select the id of every element of the permits array.
```
{
"meta": {
"source": "registry export",
"generated": "2022-01-16"
},
"permits": [
{
"id": 102,
"permit_type": "building",
"year": 2016,
"status": "pending"
... | jsonpath_write | stable | deterministic | q|jsonpath|20261022|2632 | 1,501 | 540 | 6592bc1ad4faa78e91f9dc2187671cef7c2c6228c79f4f8bcea04a9ccbffcc26 | {"answer_form": "code", "data_form": "json", "dialect": "JSONPath (RFC 9535)", "frame_id": "jsonpathw:0", "gold": {"query": "$.permits[*].id", "result": [102, 103, 107, 109, 114, 117, 120]}, "marker": "wrap:6", "presentation": "chat", "source": "JSONPath (RFC 9535)", "stratum": "JSONPath (RFC 9535):wild", "template": "... |
queries-8af137edcf26ab19 | <|start|><|user|>
jq 1.8, compact mode: show the output of the filter below.
Text:
{
"meta": {
"source": "audit sample",
"generated": "2020-09-17"
},
"permits": [
{
"id": 101,
"permit_type": "fishing",
"year": 2016,
"fee": 1950,
"status": "revoked",
"note": "refile... | jq_predict | stable | deterministic | q|jq|20261022|4267 | 1,809 | 663 | 89c2dcc3bc7d4ae6dd63bae91584abd1aff7c78801939c743283f8faf5782212 | {"answer_form": "structured", "data_form": "json", "dialect": "jq 1.8", "frame_id": "jqp:4", "gold": {"query": ".agencies[] | select(.dept == \"commerce\") | .name", "result": ["Ridgeline Trust"]}, "marker": "wrap:2", "presentation": "chat", "source": "jq 1.8", "stratum": "jq 1.8:parents_grp", "template": "parents_grp"... |
queries-7d0867e6a83ac9e8 | Given these tables, what rows does the following SQLite query produce?
Excerpt:
CREATE TABLE agencies (id INTEGER PRIMARY KEY, name TEXT NOT NULL, dept TEXT);
CREATE TABLE permits (id INTEGER PRIMARY KEY, agency_id INTEGER REFERENCES agencies(id), permit_type TEXT, year INTEGER, fee INTEGER, status TEXT);
INSERT INTO ... | sql_predict | stable | deterministic | q|sql|20261022|3707 | 1,330 | 568 | e247d7dc376ff6fc8de9f84ca4ae307378edeecb1224454d4671630ef6bed4ea | {"answer_form": "structured", "data_form": "ddl", "dialect": "SQLite 3", "frame_id": "sqlp:3", "gold": {"columns": ["id", "permit_type"], "result": [[106, "export"], [112, "export"], [120, "drilling"]], "sql": "SELECT id, permit_type FROM permits WHERE fee > 1450 ORDER BY id;"}, "marker": "wrap:5", "presentation": "pla... |
queries-b884b7f3f918b5ac | <|start|><|user|>
Apply the JSONPath expression (RFC 9535) to the document. Which values are selected?
---
{
"meta": {
"source": "registry export",
"generated": "2022-07-18"
},
"loans": [
{
"id": 101,
"purpose": "mortgage",
"year": 2015,
"status": "current"
},
{
... | jsonpath_predict | stable | deterministic | q|jsonpath|20261022|5063 | 1,813 | 649 | a6561acf85f659400674938e4f3bb0c9cf1ead804d821f048c345ea256e4a822 | {"answer_form": "structured", "data_form": "json", "dialect": "JSONPath (RFC 9535)", "frame_id": "jsonpathp:3", "gold": {"query": "$.loans[?@.principal > 800].id", "result": [107, 109, 114, 115, 121, 124]}, "marker": "wrap:1", "presentation": "chat", "source": "JSONPath (RFC 9535)", "stratum": "JSONPath (RFC 9535):filt... |
queries-6cfad7e89b656754 | <|start|><|user|>
Using the schema and data below (SQLite 3), write a SQL query that returns exactly what is asked.
---
CREATE TABLE customers (id INTEGER PRIMARY KEY, name TEXT NOT NULL, region TEXT);
CREATE TABLE invoices (id INTEGER PRIMARY KEY, customer_id INTEGER REFERENCES customers(id), product TEXT, year INTEG... | sql_write | stable | deterministic | q|sql|20261022|1011 | 1,242 | 537 | c448425cfb36ecb86a07dc8b12f447d2d1f3492c1a06b7ba7aa117b5af4e1906 | {"answer_form": "code", "data_form": "ddl", "dialect": "SQLite 3", "frame_id": "sqlw:1", "gold": {"result": [[550, 1500, 1130.0]], "sql": "SELECT MIN(amount) AS lo, MAX(amount) AS hi, ROUND(AVG(amount), 2) AS mean FROM invoices WHERE status = 'disputed';"}, "marker": "wrap:1", "presentation": "chat", "source": "SQLite ... |
queries-f3f13feffc77b5a7 | Evaluate the jq filter on the input. Print each output on its own line, as jq -c would.
---
{"meta": {"source": "batch 7", "generated": "2022-01-19"}, "invoices": [{"id": 100, "product": "hardware", "year": 2024, "amount": 1900, "status": "void", "note": "late filing"}, {"id": 104, "product": "hosting", "year": 2018, ... | jq_predict | stable | deterministic | q|jq|20261022|2609 | 1,080 | 437 | 5154a329440db8cd18b229a954c3d15c2eb832bfd573c5150802ea5c4459eeb7 | {"answer_form": "sentence", "data_form": "json", "dialect": "jq 1.8", "frame_id": "jqp:5", "gold": {"query": "[.invoices[] | .product] | unique", "result": [["hardware", "hosting"]]}, "marker": "wrap:1", "presentation": "plain", "source": "jq 1.8", "stratum": "jq 1.8:unique", "template": "unique"} |
queries-543f622b1100e80d | Predict the output of running jq -c with the filter below on this JSON (jq 1.8).
Passage:
{"meta": {"source": "registry export", "generated": "2022-03-16"}, "loans": [{"id": 102, "purpose": "student", "year": 2016, "principal": 2100}, {"id": 105, "purpose": "student", "year": 2023, "principal": 1700, "status": "defaul... | jq_predict | stable | deterministic | q|jq|20261022|8698 | 1,307 | 540 | 3504df7dd263477c315c37a6ceb06eb3f050ab90c7dbb5dcfb0dc97ce73abbc7 | {"answer_form": "sentence", "data_form": "json", "dialect": "jq 1.8", "frame_id": "jqp:1", "gold": {"query": ".lenders[] | select(.charter == \"state\") | .name", "result": ["Meridian Trust", "Ironside Logistics"]}, "marker": "wrap:3", "presentation": "plain", "source": "jq 1.8", "stratum": "jq 1.8:parents_grp", "templ... |
queries-859f17f301660a8a | Q: Write a JSONPath query (RFC 9535) that will select the nature of each cases element whose disposition equals "dismissed".
Excerpt:
{
"meta": {
"source": "quarterly extract",
"generated": "2024-06-15"
},
"cases": [
{
"id": 101,
"nature": "tort",
"year": 2018,
"damages": 850,... | jsonpath_write | stable | deterministic | q|jsonpath|20261022|5026 | 1,877 | 690 | 1d87407a5a8a07e7737e94e12294e9dce16de5b496d6d695aee8ba8e8c725645 | {"answer_form": "code", "data_form": "json", "dialect": "JSONPath (RFC 9535)", "frame_id": "jsonpathw:0", "gold": {"query": "$.cases[?@.disposition == 'dismissed'].nature", "result": ["tort", "tort", "tort", "patent", "tort", "patent"]}, "marker": "wrap:5|Q:", "presentation": "qa", "source": "JSONPath (RFC 9535)", "str... |
queries-b7c67de6d2124412 | JSONPath (RFC 9535 syntax): write the query that will select the permit_type of element 3 (zero-based) of permits in the document below.
```
{
"meta": {
"source": "audit sample",
"generated": "2025-07-10"
},
"permits": [
{
"id": 102,
"permit_type": "building",
"year": 2023,
"f... | jsonpath_write | stable | deterministic | q|jsonpath|20261022|8703 | 1,692 | 633 | c35865ba90512e1a0cd9f3a5fa824ae83e3cb2f7f94290063b90a4d683727d44 | {"answer_form": "sentence", "data_form": "json", "dialect": "JSONPath (RFC 9535)", "frame_id": "jsonpathw:3", "gold": {"query": "$.permits[3].permit_type", "result": ["building"]}, "marker": "wrap:6", "presentation": "plain", "source": "JSONPath (RFC 9535)", "stratum": "JSONPath (RFC 9535):index", "template": "index"} |
queries-e8337bac0222c515 | <|start|><|user|>
Using the schema and data below (SQLite 3), write a SQL query that returns exactly what is asked.
Passage:
CREATE TABLE lenders (id INTEGER PRIMARY KEY, name TEXT NOT NULL, charter TEXT);
CREATE TABLE loans (id INTEGER PRIMARY KEY, lender_id INTEGER REFERENCES lenders(id), purpose TEXT, year INTEGER,... | sql_write | stable | deterministic | q|sql|20261022|7525 | 984 | 395 | 6ea875e5d5ecb61075cb96730d1a6de01b4b747c74a9a9f6df8b702ba67ed5ad | {"answer_form": "code", "data_form": "ddl", "dialect": "SQLite 3", "frame_id": "sqlw:1", "gold": {"result": [["auto"]], "sql": "SELECT DISTINCT purpose FROM loans WHERE year = 2021 ORDER BY purpose;"}, "marker": "wrap:3", "presentation": "chat", "source": "SQLite 3", "stratum": "SQLite 3:distinct", "template": "distinc... |
queries-cca2a2f1992f7f22 | <|start|><|user|>
Write SQL (SQLite 3) for this report request.
<text>
Table `customers`:
| id | name | region |
|---|---|---|
| 1 | Northfield Works | north |
| 2 | Willow Trust | south |
| 3 | Harlow Supply | east |
| 4 | Oakmont Associates | south |
| 5 | Ironside Labs | north |
Table `invoices`:
| id | customer... | sql_write | stable | deterministic | q|sql|20261022|7557 | 974 | 394 | d92b4009b85379d44af1358bfbd3284c261aceca3bc85d6542b2113bdccd2ab8 | {"answer_form": "code", "data_form": "md", "dialect": "SQLite 3", "frame_id": "sqlw:6", "gold": {"result": [[8]], "sql": "SELECT COUNT(*) AS n FROM invoices WHERE status <> 'unpaid';"}, "marker": "wrap:4", "presentation": "chat", "source": "SQLite 3", "stratum": "SQLite 3:null_ne", "template": "null_ne"} |
queries-2d5f1e8d0b60d079 | Task: Write the jq program for this task: output the name of each entry in .lenders whose charter is "state".
```
{"meta": {"source": "batch 7", "generated": "2023-08-10"}, "loans": [{"id": 102, "purpose": "small_business", "year": 2022, "principal": 200, "status": "late"}, {"id": 103, "purpose": "small_business", "ye... | jq_write | stable | deterministic | q|jq|20261022|6298 | 1,182 | 463 | 69f647ec327b2ac1ed2816d38439ad8026ad6ce618d319b20c73d62b18bf9071 | {"answer_form": "code", "data_form": "json", "dialect": "jq 1.8", "frame_id": "jqw:3", "gold": {"query": ".lenders[] | select(.charter == \"state\") | .name", "result": ["Dunmore Labs", "Harlow Supply"]}, "marker": "wrap:6|Task:", "presentation": "qa", "source": "jq 1.8", "stratum": "jq 1.8:parents_grp", "template": "p... |
queries-1fab56740ad64dbb | Run the XPath 1.0 expression below on the XML and report its result.
```
<registry source="annual export">
<group code="G1" name="Ridgeline Partners">
<item id="R105" status="awarded">
<type>research</type>
<year>2021</year>
<fee>2150</fee>
</item>
<item id="R108" status="awarded">
... | xpath_predict | stable | deterministic | q|xpath|20261022|9147 | 1,643 | 654 | 580bf9e88c24bffafdddbe5579d449006ccc9dc9fc8903e30389a121dcc3be22 | {"answer_form": "sentence", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpp:5", "gold": {"query": "count(/registry/group[item])", "result": "5"}, "marker": "wrap:6", "presentation": "plain", "source": "XPath 1.0", "stratum": "XPath 1.0:count_groups", "template": "count_groups"} |
queries-98a676e69211a05f | Write a SQLite 3 query for the request below, using the tables shown.
---
CREATE TABLE vendors (id INTEGER PRIMARY KEY, name TEXT NOT NULL, tier TEXT);
CREATE TABLE contracts (id INTEGER PRIMARY KEY, vendor_id INTEGER REFERENCES vendors(id), category TEXT, year INTEGER, value INTEGER, status TEXT);
INSERT INTO vendors... | sql_write | stable | deterministic | q|sql|20261022|301 | 1,215 | 524 | 6e3b704d908bbcc69fddcb7ac211059d26b09a2c45426ec8114b677ceb591be2 | {"answer_form": "code", "data_form": "ddl", "dialect": "SQLite 3", "frame_id": "sqlw:0", "gold": {"result": [[650, 2150, 1480.0]], "sql": "SELECT MIN(value) AS lo, MAX(value) AS hi, ROUND(AVG(value), 2) AS mean FROM contracts WHERE status = 'under_review';"}, "marker": "wrap:1", "presentation": "plain", "source": "SQLi... |
queries-f1c347d6fa8a2d1b | <|start|><|user|>
Run this jq filter (jq 1.8, compact output) mentally on the document. What is printed?
---
{"meta": {"source": "quarterly extract", "generated": "2023-08-16"}, "contracts": [{"id": 100, "category": "construction", "year": 2017, "value": 350, "status": "awarded", "note": "under audit"}, {"id": 105, "c... | jq_predict | stable | deterministic | q|jq|20261022|5100 | 1,146 | 435 | 54f02d0e1ab3a18031bd0552a682b2a6b5c643966478e4065b545bdadcb50042 | {"answer_form": "structured", "data_form": "json", "dialect": "jq 1.8", "frame_id": "jqp:2", "gold": {"query": "[.contracts[] | .category] | unique", "result": [["construction", "research"]]}, "marker": "wrap:1", "presentation": "chat", "source": "jq 1.8", "stratum": "jq 1.8:unique", "template": "unique"} |
queries-946eefef2583f0fb | Request: Run the RFC 9535 JSONPath query on the input and report the selected values as a JSON array.
<text>
{
"meta": {
"source": "quarterly extract",
"generated": "2020-09-18"
},
"contracts": [
{
"id": 101,
"category": "consulting",
"year": 2021,
"value": 1200,
"status... | jsonpath_predict | stable | deterministic | q|jsonpath|20261022|8573 | 2,094 | 767 | 32d6832bf7df6817a79ac4a3df45adb0d48b3d399bb2124137de1d138628bdc8 | {"answer_form": "structured", "data_form": "json", "dialect": "JSONPath (RFC 9535)", "frame_id": "jsonpathp:5", "gold": {"query": "$.contracts[*].id", "result": [101, 105, 108, 111, 112, 117, 119, 123, 126, 127]}, "marker": "wrap:4|Request:", "presentation": "qa", "source": "JSONPath (RFC 9535)", "stratum": "JSONPath (... |
queries-8e9cce2622789db4 | ### Task
Write an XPath 1.0 expression that returns the id of the last item in the third group, as a string (empty if that group has no items).
"""
<registry source="annual export">
<group code="G1" name="Fairhaven Works">
<item id="R100" status="pending">
<type>tort</type>
<year>2018</year>
<f... | xpath_write | stable | deterministic | q|xpath|20261022|5919 | 1,700 | 691 | 099dd1d957d0a4e331d5f32d5daaedfc065abbdbb65e5235a0f5e33dcec3d3af | {"answer_form": "sentence", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpw:0", "gold": {"query": "string(/registry/group[3]/item[last()]/@id)", "result": "R128"}, "marker": "wrap:0|### Task", "presentation": "qa", "source": "XPath 1.0", "stratum": "XPath 1.0:positional", "template": "positional"} |
queries-9e7a0d9ab16c092c | Request: Run this SQLite 3 query in your head against the tables below. What does it return?
Passage:
CREATE TABLE agencies (id INTEGER PRIMARY KEY, name TEXT NOT NULL, dept TEXT);
CREATE TABLE permits (id INTEGER PRIMARY KEY, agency_id INTEGER REFERENCES agencies(id), permit_type TEXT, year INTEGER, fee INTEGER, stat... | sql_predict | stable | deterministic | q|sql|20261022|7797 | 1,606 | 709 | 8738d56e623f88d9de0dd445209d90561cefc56da780d93231eb41781a8b034b | {"answer_form": "sentence", "data_form": "ddl", "dialect": "SQLite 3", "frame_id": "sqlp:0", "gold": {"columns": ["lo", "hi", "mean"], "result": [[1200, 2350, 1800.0]], "sql": "SELECT MIN(fee) AS lo, MAX(fee) AS hi, ROUND(AVG(fee), 2) AS mean FROM permits WHERE status = 'revoked';"}, "marker": "wrap:3|Request:", "prese... |
queries-1fd14a3446dd49ce | Q: jq 1.8: compose a filter over the input JSON that will collect into an array the ids of records in .loans that have a note key.
---
{"meta": {"source": "batch 7", "generated": "2024-06-18"}, "loans": [{"id": 102, "purpose": "auto", "year": 2018, "principal": 1150, "status": "current"}, {"id": 103, "purpose": "auto"... | jq_write | stable | deterministic | q|jq|20261022|3201 | 1,058 | 451 | 6780e89403541bce50a7c1a6ea37a506ffa2efa47a0d8800395c6b6d96fb4978 | {"answer_form": "sentence", "data_form": "json", "dialect": "jq 1.8", "frame_id": "jqw:4", "gold": {"query": "[.loans[] | select(has(\"note\")) | .id]", "result": [[108, 117]]}, "marker": "wrap:1|Q:", "presentation": "qa", "source": "jq 1.8", "stratum": "jq 1.8:has_note", "template": "has_note"} |
queries-d0107a89a0d90616 | Evaluate the jq filter on the input. Print each output on its own line, as jq -c would.
---
{"meta": {"source": "audit sample", "generated": "2020-03-12"}, "invoices": [{"id": 101, "product": "support", "year": 2021, "amount": 1450, "status": "unpaid"}, {"id": 103, "product": "training", "year": 2018, "amount": 1500, ... | jq_predict | stable | deterministic | q|jq|20261022|7552 | 936 | 397 | efbab12082b5d7f8e95bcd00a7ca07ed87deeca97b57fa13cef5b654a85e5153 | {"answer_form": "sentence", "data_form": "json", "dialect": "jq 1.8", "frame_id": "jqp:5", "gold": {"query": "[.invoices[] | .amount // 0] | add", "result": [8600]}, "marker": "wrap:1", "presentation": "plain", "source": "jq 1.8", "stratum": "jq 1.8:sum_alt", "template": "sum_alt"} |
queries-822a72b448007993 | <|start|><|user|>
Dialect: SQLite 3. Show the result set of this query.
```
CREATE TABLE parties (id INTEGER PRIMARY KEY, name TEXT NOT NULL, kind TEXT);
CREATE TABLE filings (id INTEGER PRIMARY KEY, party_id INTEGER REFERENCES parties(id), court TEXT, year INTEGER, fee INTEGER, status TEXT);
INSERT INTO parties VALUE... | sql_predict | stable | deterministic | q|sql|20261022|9270 | 1,165 | 542 | 8623f5140de8f2b66bb4fcb658a3f32fa6c1a9007c26253c322abefed2a35ce1 | {"answer_form": "sentence", "data_form": "ddl", "dialect": "SQLite 3", "frame_id": "sqlp:5", "gold": {"columns": ["court", "n"], "result": [["appeals", 5], ["bankruptcy", 3]], "sql": "SELECT court, COUNT(*) AS n FROM filings GROUP BY court HAVING COUNT(*) >= 3 ORDER BY n DESC, court;"}, "marker": "wrap:6", "presentatio... |
queries-96830ef0ee742970 | Given these tables, what rows does the following SQLite query produce?
```
Table `agencies`:
| id | name | dept |
|---|---|---|
| 1 | Alder Associates | labor |
| 2 | Zephyr Services | energy |
| 3 | Fairhaven Labs | labor |
| 4 | Meridian Group | energy |
Table `permits`:
| id | agency_id | permit_type | year | fe... | sql_predict | stable | deterministic | q|sql|20261022|5684 | 839 | 375 | 5b18e01b796feaf1dadb0ba4760d142f16796b80d50281db5d5ca44b19fed3d1 | {"answer_form": "structured", "data_form": "md", "dialect": "SQLite 3", "frame_id": "sqlp:3", "gold": {"columns": ["permit_type", "n", "total"], "result": [["building", 4, 7450], ["drilling", 3, 4200]], "sql": "SELECT permit_type, COUNT(*) AS n, SUM(fee) AS total FROM permits GROUP BY permit_type ORDER BY permit_type;"... |
queries-bd8cd5260e2fa2be | [Task]
Write one RFC 9535 JSONPath expression to select the id of the last element of invoices.
```
{"meta": {"source": "quarterly extract", "generated": "2021-05-13"}, "invoices": [{"id": 100, "product": "training", "year": 2021, "amount": 1250, "status": "paid"}, {"id": 105, "product": "training", "year": 2020, "amo... | jsonpath_write | stable | deterministic | q|jsonpath|20261022|821 | 1,122 | 441 | 9d7ba8c473deba9b436c3491c611ba4047e6550ace8595794e246c677545aba1 | {"answer_form": "code", "data_form": "json", "dialect": "JSONPath (RFC 9535)", "frame_id": "jsonpathw:5", "gold": {"query": "$.invoices[-1].id", "result": [119]}, "marker": "wrap:6|[Task]", "presentation": "qa", "source": "JSONPath (RFC 9535)", "stratum": "JSONPath (RFC 9535):neg_index", "template": "neg_index"} |
queries-fe92022013193ab8 | <|start|><|user|>
Dialect: SQLite 3. Write the SELECT statement that answers the request.
"""
Table `customers`:
| id | name | region |
|---|---|---|
| 1 | Larkspur Associates | south |
| 2 | Harlow Associates | west |
| 3 | Zephyr Labs | west |
| 4 | Thistle Services | west |
| 5 | Upland Partners | west |
| 6 | Van... | sql_write | stable | deterministic | q|sql|20261022|400 | 1,012 | 432 | a381f92275461410cd9978e4fa6b9cbc14c161c8b30dcf2698ca378b3a55d95f | {"answer_form": "sentence", "data_form": "md", "dialect": "SQLite 3", "frame_id": "sqlw:2", "gold": {"result": [[100, "high"], [105, "low"], [108, "low"], [110, "low"], [112, "high"], [116, "high"], [118, "unknown"], [121, "low"]], "sql": "SELECT id, CASE WHEN amount IS NULL THEN 'unknown' WHEN amount >= 1700 THEN 'hig... |
queries-bd0be721dc254fdb | Evaluate the RFC 9535 JSONPath query below and list the selected values as a JSON array.
```
{
"meta": {
"source": "registry export",
"generated": "2024-03-17"
},
"permits": [
{
"id": 101,
"permit_type": "drilling",
"year": 2016,
"fee": 2100,
"status": "pending",
"... | jsonpath_predict | stable | deterministic | q|jsonpath|20261022|9761 | 1,572 | 570 | 6a4a01421cc2e1f07a6e4ddc009fd0da44f15d00b15c90a9d46c5f11614aed49 | {"answer_form": "structured", "data_form": "json", "dialect": "JSONPath (RFC 9535)", "frame_id": "jsonpathp:1", "gold": {"query": "$.permits[?@.note].id", "result": [101, 107]}, "marker": "wrap:6", "presentation": "plain", "source": "JSONPath (RFC 9535)", "stratum": "JSONPath (RFC 9535):exists", "template": "exists"} |
queries-68f7d67cb5bcf839 | XPath 1.0 semantics: what value or node-set does this expression produce?
<text>
<registry source="field office">
<group code="G1" name="Dunmore Services">
<item id="R103" status="paid_off">
<type>small_business</type>
<year>2019</year>
<fee>2400</fee>
</item>
</group>
<group code="G2" ... | xpath_predict | stable | deterministic | q|xpath|20261022|3598 | 1,853 | 739 | 603fe427ca919fbd1d0dc44b16a0938bbbb33aa8cdd435c2f87b276ab4e8daa8 | {"answer_form": "bare", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpp:4", "gold": {"query": "sum(//group[@code='G4']/item/fee)", "result": "4450"}, "marker": "wrap:4", "presentation": "plain", "source": "XPath 1.0", "stratum": "XPath 1.0:sum_group", "template": "sum_group"} |
queries-361a4ca50731a36b | Evaluate the SQL query (SQLite 3 semantics, including NULL handling) on the data below.
Text:
CREATE TABLE vendors (id INTEGER PRIMARY KEY, name TEXT NOT NULL, tier TEXT);
CREATE TABLE contracts (id INTEGER PRIMARY KEY, vendor_id INTEGER REFERENCES vendors(id), category TEXT, year INTEGER, value INTEGER, status TEXT);... | sql_predict | stable | deterministic | q|sql|20261022|402 | 1,535 | 645 | 86f07670b7f783f11e13583db4e07f52c712525f232bf0ce9d935dd093a14ab6 | {"answer_form": "structured", "data_form": "ddl", "dialect": "SQLite 3", "frame_id": "sqlp:4", "gold": {"columns": ["id"], "result": [[115], [123], [132]], "sql": "SELECT id FROM contracts WHERE value > (SELECT AVG(value) FROM contracts) ORDER BY id;"}, "marker": "wrap:2", "presentation": "plain", "source": "SQLite 3",... |
queries-160cfb476e4b5898 | <|start|><|user|>
Using XPath 1.0, write an expression for the total of the fee values of the items in the group with code 'G1' in the XML below.
<text>
<registry source="field office">
<group code="G1" name="Pinecrest Holdings">
<item id="R105" status="stayed">
<type>claims</type>
<year>2021</year>
... | xpath_write | stable | deterministic | q|xpath|20261022|6565 | 1,259 | 488 | d57b81b5177b8ffc16012f7e6a78070623ef55d10938c03dfab5ac4f6879c25c | {"answer_form": "code", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpw:1", "gold": {"query": "sum(//group[@code='G1']/item/fee)", "result": "4300"}, "marker": "wrap:4", "presentation": "chat", "source": "XPath 1.0", "stratum": "XPath 1.0:sum_group", "template": "sum_group"} |
queries-a04bc3e554b8dad2 | <|start|><|user|>
Predict the output of the query (SQLite 3) for the tables shown.
<text>
Table `lenders`:
| id | name | charter |
|---|---|---|
| 1 | Oakmont Trust | state |
| 2 | Pinecrest Holdings | state |
| 3 | Dunmore Group | state |
| 4 | Ridgeline Works | national |
Table `loans`:
| id | lender_id | purpose... | sql_predict | stable | deterministic | q|sql|20261022|2132 | 803 | 334 | 63fe0c0256ed72b2c36ff2570bac38808ca9e92b0e6e68971bc0d1611f857844 | {"answer_form": "structured", "data_form": "md", "dialect": "SQLite 3", "frame_id": "sqlp:2", "gold": {"columns": ["id", "principal"], "result": [[116, 1900], [108, 1850], [110, 1300]], "sql": "SELECT id, principal FROM loans WHERE principal IS NOT NULL ORDER BY principal DESC, id LIMIT 3;"}, "marker": "wrap:4", "prese... |
queries-ff12f205dd465525 | Q: Compose a JSONPath query (RFC 9535) for this: select every value value anywhere in the document.
Excerpt:
{
"meta": {
"source": "registry export",
"generated": "2022-02-10"
},
"contracts": [
{
"id": 102,
"category": "supplies",
"year": 2021,
"value": 650,
"status": "t... | jsonpath_write | stable | deterministic | q|jsonpath|20261022|7123 | 1,951 | 684 | a60ecee34e656292cf24a18d8367a61cde1561df7ba0c1820d98fcf004b0ae84 | {"answer_form": "code", "data_form": "json", "dialect": "JSONPath (RFC 9535)", "frame_id": "jsonpathw:4", "gold": {"query": "$..value", "result": [650, 1450, 1600, 1350, 1750, 2000, 1400]}, "marker": "wrap:5|Q:", "presentation": "qa", "source": "JSONPath (RFC 9535)", "stratum": "JSONPath (RFC 9535):descendant", "templa... |
queries-12749bea210a6628 | <|start|><|user|>
What nodes does this JSONPath query (RFC 9535) select from the document? Give their values as a JSON array.
```
{
"meta": {
"source": "quarterly extract",
"generated": "2025-01-15"
},
"invoices": [
{
"id": 100,
"product": "training",
"year": 2021,
"amount": 1... | jsonpath_predict | stable | deterministic | q|jsonpath|20261022|7641 | 1,331 | 483 | 7d00461308c34f11cbd38dcc690972a9002948590cf263459dee12de7d6c36a0 | {"answer_form": "sentence", "data_form": "json", "dialect": "JSONPath (RFC 9535)", "frame_id": "jsonpathp:0", "gold": {"query": "$.invoices[-1].id", "result": [115]}, "marker": "wrap:6", "presentation": "chat", "source": "JSONPath (RFC 9535)", "stratum": "JSONPath (RFC 9535):neg_index", "template": "neg_index"} |
queries-b51fee5aaebccd94 | <|start|><|user|>
Using the schema and data below (SQLite 3), write a SQL query that returns exactly what is asked.
```
CREATE TABLE agencies (id INTEGER PRIMARY KEY, name TEXT NOT NULL, dept TEXT);
CREATE TABLE permits (id INTEGER PRIMARY KEY, agency_id INTEGER REFERENCES agencies(id), permit_type TEXT, year INTEGER,... | sql_write | stable | deterministic | q|sql|20261022|9628 | 1,064 | 438 | 909d5a7ba4701a88537fe924bba3512b7be75a57c044d38fdaca63704f471864 | {"answer_form": "code", "data_form": "ddl", "dialect": "SQLite 3", "frame_id": "sqlw:1", "gold": {"result": [[104, "export"], [108, "export"], [112, "export"], [123, "grazing"]], "sql": "SELECT id, permit_type FROM permits WHERE fee > 1400 ORDER BY id;"}, "marker": "wrap:6", "presentation": "chat", "source": "SQLite 3"... |
queries-3c34e1beca6c272f | Question: Evaluate the SQL query (SQLite 3 semantics, including NULL handling) on the data below.
---
Table `vendors`:
| id | name | tier |
|---|---|---|
| 1 | Kestrel Associates | silver |
| 2 | Ironside Group | gold |
| 3 | Cobalt Logistics | bronze |
| 4 | Quarry Group | bronze |
Table `contracts`:
| id | vendor... | sql_predict | stable | deterministic | q|sql|20261022|5131 | 944 | 353 | 88359091871272484c5b130c2f4c7adcd5fd3c999d8c7a4328eed620dbdb054d | {"answer_form": "structured", "data_form": "md", "dialect": "SQLite 3", "frame_id": "sqlp:4", "gold": {"columns": ["category", "n"], "result": [["maintenance", 4]], "sql": "SELECT category, COUNT(*) AS n FROM contracts GROUP BY category HAVING COUNT(*) >= 3 ORDER BY n DESC, category;"}, "marker": "wrap:1|Question:", "p... |
queries-e3225ed9706d62cf | JSONPath, RFC 9535 semantics: what is the nodelist (as an array of values)?
Excerpt:
{
"meta": {
"source": "batch 7",
"generated": "2021-08-16"
},
"contracts": [
{
"id": 100,
"category": "maintenance",
"year": 2021,
"value": 1350,
"status": "under_review"
},
{
... | jsonpath_predict | stable | deterministic | q|jsonpath|20261022|3646 | 1,481 | 543 | bac656520ba8be78f7e702b040259fba37c7f0074c62a3a1dbe61f4dd836bdd2 | {"answer_form": "sentence", "data_form": "json", "dialect": "JSONPath (RFC 9535)", "frame_id": "jsonpathp:4", "gold": {"query": "$.contracts[?@.value > 2100].id", "result": []}, "marker": "wrap:5", "presentation": "plain", "source": "JSONPath (RFC 9535)", "stratum": "JSONPath (RFC 9535):filter_gt", "template": "filter_... |
queries-c2719bdb4d4365bd | Apply the JSONPath expression (RFC 9535) to the document. Which values are selected?
Text:
{
"meta": {
"source": "quarterly extract",
"generated": "2020-08-18"
},
"invoices": [
{
"id": 100,
"product": "hardware",
"year": 2019,
"amount": 1850,
"status": "paid"
},
... | jsonpath_predict | stable | deterministic | q|jsonpath|20261022|9658 | 1,904 | 708 | bedaef0da8da3b277c6510fa385442776a58caf239b6737713538e8a095a0c83 | {"answer_form": "sentence", "data_form": "json", "dialect": "JSONPath (RFC 9535)", "frame_id": "jsonpathp:3", "gold": {"query": "$.invoices[?@.amount > 1500].id", "result": [100, 109, 121]}, "marker": "wrap:2", "presentation": "plain", "source": "JSONPath (RFC 9535)", "stratum": "JSONPath (RFC 9535):filter_gt", "templa... |
queries-12e4ec4ad142918a | Evaluate the jq filter on the input. Print each output on its own line, as jq -c would.
<text>
{"meta": {"source": "audit sample", "generated": "2021-06-15"}, "filings": [{"id": 102, "court": "appeals", "year": 2022, "fee": 1350, "status": "stayed", "note": "amended"}, {"id": 103, "court": "appeals", "year": 2015, "fe... | jq_predict | stable | deterministic | q|jq|20261022|3022 | 1,499 | 624 | e1e4b2407a52041697184e487bce002ef61ee6ee3cddc21a549b16aebc66f9c8 | {"answer_form": "sentence", "data_form": "json", "dialect": "jq 1.8", "frame_id": "jqp:5", "gold": {"query": ".filings | map(select(.status == \"stayed\")) | length", "result": [6]}, "marker": "wrap:4", "presentation": "plain", "source": "jq 1.8", "stratum": "jq 1.8:count_status", "template": "count_status"} |
queries-3e08c3ecb476f504 | Q: Dialect: SQLite 3. Write the SELECT statement that answers the request.
---
Table `courts`:
| id | name | circuit |
|---|---|---|
| 1 | Zephyr Trust | second |
| 2 | Northfield Group | federal |
| 3 | Ironside Trust | first |
| 4 | Alder Labs | ninth |
Table `cases`:
| id | court_id | nature | year | damages | d... | sql_write | stable | deterministic | q|sql|20261022|2542 | 953 | 382 | 9dff693d7154b3c8b793200e5ec422152be607cef8bb7e60c86d720d49b6d478 | {"answer_form": "code", "data_form": "md", "dialect": "SQLite 3", "frame_id": "sqlw:2", "gold": {"result": [[3]], "sql": "SELECT COUNT(*) AS n FROM cases WHERE disposition <> 'judgment';"}, "marker": "wrap:1|Q:", "presentation": "qa", "source": "SQLite 3", "stratum": "SQLite 3:null_ne", "template": "null_ne"} |
queries-6c80a3f258bd3138 | <|start|><|user|>
Dialect: SQLite 3. Show the result set of this query.
Passage:
Table `vendors`:
| id | name | tier |
|---|---|---|
| 1 | Dunmore Holdings | gold |
| 2 | Yarrow Group | bronze |
| 3 | Fairhaven Services | silver |
Table `contracts`:
| id | vendor_id | category | year | value | status |
|---|---|---... | sql_predict | stable | deterministic | q|sql|20261022|8388 | 1,024 | 489 | b7369a774c46abb2b49615c557cacd0e9d93e5fd80a04f7c5b99b3cf653283b6 | {"answer_form": "structured", "data_form": "md", "dialect": "SQLite 3", "frame_id": "sqlp:5", "gold": {"columns": ["id", "category"], "result": [[106, "consulting"], [111, "supplies"], [115, "consulting"], [120, "research"], [123, "construction"], [126, "supplies"], [128, "construction"], [130, "supplies"]], "sql": "SE... |
queries-9e97a30c9066b7a7 | Here are two SQLite tables. Write the query for the question that follows.
"""
Table `parties`:
| id | name | kind |
|---|---|---|
| 1 | Upland Partners | individual |
| 2 | Willow Works | corporation |
| 3 | Ironside Supply | nonprofit |
| 4 | Harlow Partners | individual |
| 5 | Alder Labs | agency |
Table `filing... | sql_write | stable | deterministic | q|sql|20261022|3378 | 874 | 347 | 4cdf097bae74d5bfc620428edfd6cd1512023b339a4eac5c083534f0c7549975 | {"answer_form": "code", "data_form": "md", "dialect": "SQLite 3", "frame_id": "sqlw:5", "gold": {"result": [[104, "appeals"], [106, "claims"], [119, "claims"]], "sql": "SELECT id, court FROM filings WHERE fee > 1900 ORDER BY id;"}, "marker": "wrap:0", "presentation": "plain", "source": "SQLite 3", "stratum": "SQLite 3:... |
queries-03be4bcb45adef29 | <|start|><|user|>
Execute the query mentally and give its output. Database: SQLite 3.
Passage:
CREATE TABLE vendors (id INTEGER PRIMARY KEY, name TEXT NOT NULL, tier TEXT);
CREATE TABLE contracts (id INTEGER PRIMARY KEY, vendor_id INTEGER REFERENCES vendors(id), category TEXT, year INTEGER, value INTEGER, status TEXT)... | sql_predict | stable | deterministic | q|sql|20261022|5261 | 917 | 377 | 01cf8b01f94aa1413e3a2d884c11741ed07d349fca9b4c164829e1ae338be1a0 | {"answer_form": "sentence", "data_form": "ddl", "dialect": "SQLite 3", "frame_id": "sqlp:6", "gold": {"columns": ["id", "category"], "result": [], "sql": "SELECT id, category FROM contracts WHERE value > 1950 ORDER BY id;"}, "marker": "wrap:3", "presentation": "chat", "source": "SQLite 3", "stratum": "SQLite 3:filter_g... |
queries-fe1330c1d891e0b9 | <|start|><|user|>
Using the schema and data below (SQLite 3), write a SQL query that returns exactly what is asked.
Passage:
CREATE TABLE customers (id INTEGER PRIMARY KEY, name TEXT NOT NULL, region TEXT);
CREATE TABLE invoices (id INTEGER PRIMARY KEY, customer_id INTEGER REFERENCES customers(id), product TEXT, year ... | sql_write | stable | deterministic | q|sql|20261022|2862 | 1,465 | 604 | c074668a901714b9c19ec228176d3ee6c02171497aa83d35e2c67096a1fbef25 | {"answer_form": "code", "data_form": "ddl", "dialect": "SQLite 3", "frame_id": "sqlw:1", "gold": {"result": [[100], [122], [124], [128]], "sql": "SELECT id FROM invoices WHERE amount > (SELECT AVG(amount) FROM invoices) ORDER BY id;"}, "marker": "wrap:3", "presentation": "chat", "source": "SQLite 3", "stratum": "SQLite... |
queries-4a1f7ad9ba533cf0 | <|start|><|user|>
Apply the XPath 1.0 expression to the document and give the result.
Excerpt:
<registry source="field office">
<group code="G1" name="Alder Supply">
<item id="R113" status="current">
<type>student</type>
<year>2018</year>
<fee>2600</fee>
</item>
<item id="R132" status="... | xpath_predict | stable | deterministic | q|xpath|20261022|5169 | 1,801 | 724 | d512f38603d68a89dd790b32d534328a94a657a9b35319ad053753418c266f77 | {"answer_form": "sentence", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpp:3", "gold": {"query": "sum(//group[@code='G3']/item/fee)", "result": "4850"}, "marker": "wrap:5", "presentation": "chat", "source": "XPath 1.0", "stratum": "XPath 1.0:sum_group", "template": "sum_group"} |
queries-2416e9ae0baaf20b | [Task]
XPath 1.0: compose an expression evaluating to the text of the type element of each item whose status is 'stayed'.
```
<registry source="audit copy">
<group code="G1" name="Stonebrook Associates">
<item id="R108" status="dismissed" expires="2019-12-31">
<type>tax</type>
<year>2018</year>
... | xpath_write | stable | deterministic | q|xpath|20261022|3930 | 1,798 | 761 | df1e81116c79b33c48d9bef4b564ceffbaa585387116b098f38f639aff18d302 | {"answer_form": "sentence", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpw:3", "gold": {"query": "//item[@status='stayed']/type/text()", "result": ["bankruptcy", "appeals", "district", "district"]}, "marker": "wrap:6|[Task]", "presentation": "qa", "source": "XPath 1.0", "stratum": "XPath 1.0:text_values",... |
queries-c5d7c34e0f98f0c4 | <|start|><|system|>
You are a helpful assistant.<|eot|>
<|user|>
Given these tables, what rows does the following SQLite query produce?
```
Table `parties`:
| id | name | kind |
|---|---|---|
| 1 | Thistle Associates | individual |
| 2 | Fairhaven Supply | individual |
| 3 | Cobalt Group | individual |
| 4 | Thistle ... | sql_predict | stable | deterministic | q|sql|20261022|2287 | 1,103 | 398 | 49641665c80a375734cb1d53fa9d244fda7ade75e6436a6a3f858d26dbcfd85a | {"answer_form": "structured", "data_form": "md", "dialect": "SQLite 3", "frame_id": "sqlp:3", "gold": {"columns": ["name"], "result": [["Thistle Labs"]], "sql": "SELECT name FROM parties WHERE NOT EXISTS (SELECT 1 FROM filings WHERE filings.party_id = parties.id) ORDER BY id;"}, "marker": "wrap:6", "presentation": "cha... |
queries-9ef0812304dfb021 | <|start|><|user|>
Write one XPath 1.0 expression that yields the id attributes of all item elements whose status attribute is 'default'.
---
<registry source="annual export">
<group code="G1" name="Dunmore Services">
<item id="R121" status="default">
<type>personal</type>
<year>2024</year>
<fee... | xpath_write | stable | deterministic | q|xpath|20261022|639 | 1,464 | 559 | 80a7e0be3ed5bb6a4f9ed622f4ab7a78947ceb09712367e393ef89bd331a5a18 | {"answer_form": "code", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpw:4", "gold": {"query": "//item[@status='default']/@id", "result": ["R121", "R111", "R116", "R105", "R108"]}, "marker": "wrap:1", "presentation": "chat", "source": "XPath 1.0", "stratum": "XPath 1.0:attr_filter", "template": "attr_filter... |
queries-22870d47679d9099 | Using JSONPath as defined in RFC 9535, write a query to select the id of the last element of invoices.
Excerpt:
{
"meta": {
"source": "registry export",
"generated": "2020-04-17"
},
"invoices": [
{
"id": 100,
"product": "hosting",
"year": 2021,
"amount": 950,
"status": "... | jsonpath_write | stable | deterministic | q|jsonpath|20261022|8421 | 1,396 | 498 | 4afda2e1a26d17ceca1ab0d40bfea338f729f14974937e44183d9d54f9051200 | {"answer_form": "code", "data_form": "json", "dialect": "JSONPath (RFC 9535)", "frame_id": "jsonpathw:1", "gold": {"query": "$.invoices[-1].id", "result": [120]}, "marker": "wrap:5", "presentation": "plain", "source": "JSONPath (RFC 9535)", "stratum": "JSONPath (RFC 9535):neg_index", "template": "neg_index"} |
queries-06f5c4ee3023227c | <|start|><|user|>
What does the XPath 1.0 expression below return for this document?
"""
<registry source="annual export">
<group code="G1" name="Granite Trust">
<item id="R105">
<type>export</type>
<year>2020</year>
<fee>1550</fee>
</item>
<item id="R106" status="revoked">
<type>... | xpath_predict | stable | deterministic | q|xpath|20261022|3955 | 1,426 | 580 | bf6eef49d0089c3bf45c6b55ebdd084759947757826addacfa390c3dddb5580b | {"answer_form": "sentence", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpp:1", "gold": {"query": "string(/registry/group[1]/item[last()]/@id)", "result": "R113"}, "marker": "wrap:0", "presentation": "chat", "source": "XPath 1.0", "stratum": "XPath 1.0:positional", "template": "positional"} |
queries-506ebd2dcbb62d8d | Evaluate this XPath 1.0 expression against the XML document. What is the result?
"""
<registry source="audit copy">
<group code="G1" name="Larkspur Works"/>
<group code="G2" name="Meridian Works">
<item id="R122" status="paid_off">
<type>personal</type>
<year>2024</year>
</item>
<item id="R... | xpath_predict | stable | deterministic | q|xpath|20261022|4181 | 1,536 | 643 | f1a2a11182d5a705a1cbf2b408b5d3c0b5a4eaf1797e69fb0d915935c08295d4 | {"answer_form": "structured", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpp:0", "gold": {"query": "//item[not(@expires)]/@id", "result": ["R122", "R124", "R108", "R109", "R113", "R118", "R116", "R101"]}, "marker": "wrap:0", "presentation": "plain", "source": "XPath 1.0", "stratum": "XPath 1.0:not_attr", ... |
queries-7115d287a24d62d6 | Instruction: Give the XPath 1.0 query that returns the name attributes of the groups containing an item whose type is 'tax' (each group once, in document order).
---
<registry source="audit copy">
<group code="G1" name="Fairhaven Labs">
<item id="R103" status="dismissed">
<type>tax</type>
<year>2022<... | xpath_write | stable | deterministic | q|xpath|20261022|6761 | 1,315 | 512 | 37c089759bb2c821a450f6023c1fe44238c81da2e151e1b1e860ff95bf3e15cc | {"answer_form": "code", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpw:2", "gold": {"query": "//item[type='tax']/../@name", "result": ["Fairhaven Labs", "Kestrel Works", "Upland Supply"]}, "marker": "wrap:1|Instruction:", "presentation": "qa", "source": "XPath 1.0", "stratum": "XPath 1.0:parent_axis", "te... |
queries-01d31a259a5431b3 | Given these tables, what rows does the following SQLite query produce?
```
Table `courts`:
| id | name | circuit |
|---|---|---|
| 1 | Oakmont Associates | ninth |
| 2 | Upland Works | ninth |
| 3 | Eastgate Supply | second |
Table `cases`:
| id | court_id | nature | year | damages | disposition |
|---|---|---|---|... | sql_predict | stable | deterministic | q|sql|20261022|2558 | 890 | 488 | e11d53d9f633a5ca1c8a16a256aa0184cc09ad701084e6056f939537f4b4b74b | {"answer_form": "structured", "data_form": "md", "dialect": "SQLite 3", "frame_id": "sqlp:3", "gold": {"columns": ["id", "band"], "result": [[101, "low"], [104, "low"], [106, "low"], [110, "high"], [114, "unknown"], [115, "high"], [118, "high"], [122, "high"]], "sql": "SELECT id, CASE WHEN damages IS NULL THEN 'unknown... |
queries-ba4d319aa3845a86 | JSONPath (RFC 9535 syntax): write the query that will select the category of element 2 (zero-based) of contracts in the document below.
Text:
{
"meta": {
"source": "quarterly extract",
"generated": "2025-05-14"
},
"contracts": [
{
"id": 101,
"category": "construction",
"year": 2018,... | jsonpath_write | stable | deterministic | q|jsonpath|20261022|732 | 2,210 | 756 | 2d77de3b19164031de40bd20ad4f51c2987833ea72993bf17e4a4b39b1d5d776 | {"answer_form": "code", "data_form": "json", "dialect": "JSONPath (RFC 9535)", "frame_id": "jsonpathw:3", "gold": {"query": "$.contracts[2].category", "result": ["construction"]}, "marker": "wrap:2", "presentation": "plain", "source": "JSONPath (RFC 9535)", "stratum": "JSONPath (RFC 9535):index", "template": "index"} |
queries-bce5f7daf92173a5 | <|start|><|user|>
Here are two SQLite tables. Write the query for the question that follows.
Excerpt:
Table `customers`:
| id | name | region |
|---|---|---|
| 1 | Pinecrest Labs | south |
| 2 | Stonebrook Associates | north |
| 3 | Fairhaven Logistics | north |
| 4 | Larkspur Partners | west |
| 5 | Thistle Labs | s... | sql_write | stable | deterministic | q|sql|20261022|1423 | 1,052 | 419 | 6318d0531121f08987ba3fb46d6c2a006600d4fd9c80bf95d56c02807061ef42 | {"answer_form": "code", "data_form": "md", "dialect": "SQLite 3", "frame_id": "sqlw:5", "gold": {"result": [["hosting"], ["training"]], "sql": "SELECT DISTINCT product FROM invoices WHERE year = 2017 ORDER BY product;"}, "marker": "wrap:5", "presentation": "chat", "source": "SQLite 3", "stratum": "SQLite 3:distinct", "... |
queries-66243710b58a0fe2 | Give a SQLite query that produces the described result from these tables.
"""
CREATE TABLE parties (id INTEGER PRIMARY KEY, name TEXT NOT NULL, kind TEXT);
CREATE TABLE filings (id INTEGER PRIMARY KEY, party_id INTEGER REFERENCES parties(id), court TEXT, year INTEGER, fee INTEGER, status TEXT);
INSERT INTO parties VAL... | sql_write | stable | deterministic | q|sql|20261022|5197 | 1,070 | 489 | ed87d72c62232d196d70acf9d700e88c724820bc244a5f11c63744f5e0fd51cf | {"answer_form": "code", "data_form": "ddl", "dialect": "SQLite 3", "frame_id": "sqlw:3", "gold": {"result": [[105, "Vantage Group"], [108, "Vantage Group"], [115, "Meridian Associates"], [118, "Vantage Group"], [121, "Meridian Associates"], [124, "Vantage Group"]], "sql": "SELECT filings.id, parties.name FROM filings J... |
queries-c9dfe94d8c30207a | [Task]
XPath 1.0: compose an expression evaluating to the total of the fee values of the items in the group with code 'G3'.
"""
<registry source="field office">
<group code="G1" name="Juniper Group"/>
<group code="G2" name="Oakmont Trust">
<item id="R112" status="paid">
<type>hardware</type>
<year>... | xpath_write | stable | deterministic | q|xpath|20261022|3846 | 1,233 | 482 | 5064abb29d8f17de795f64979612c5f7516737a2a0f4e9fe69dbf972248f51ac | {"answer_form": "code", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpw:3", "gold": {"query": "sum(//group[@code='G3']/item/fee)", "result": "1850"}, "marker": "wrap:0|[Task]", "presentation": "qa", "source": "XPath 1.0", "stratum": "XPath 1.0:sum_group", "template": "sum_group"} |
queries-f8a669070a881c5f | ### Task
What is the result of the SQLite 3 query below on this data?
```
CREATE TABLE agencies (id INTEGER PRIMARY KEY, name TEXT NOT NULL, dept TEXT);
CREATE TABLE permits (id INTEGER PRIMARY KEY, agency_id INTEGER REFERENCES agencies(id), permit_type TEXT, year INTEGER, fee INTEGER, status TEXT);
INSERT INTO agenci... | sql_predict | stable | deterministic | q|sql|20261022|5820 | 1,012 | 421 | 2c2bdda436f773c035cdfb9808318ff4dffc4a0c4b91a2a6115e898e062c8ec8 | {"answer_form": "sentence", "data_form": "ddl", "dialect": "SQLite 3", "frame_id": "sqlp:1", "gold": {"columns": ["id"], "result": [[105], [107]], "sql": "SELECT id FROM permits WHERE status = 'expired' AND year >= 2018 ORDER BY id;"}, "marker": "wrap:6|### Task", "presentation": "qa", "source": "SQLite 3", "stratum": ... |
queries-9f7070d2c830d015 | Write one XPath 1.0 expression that yields the name attributes of the groups containing an item whose type is 'training' (each group once, in document order).
Excerpt:
<registry source="annual export">
<group code="G1" name="Fairhaven Logistics">
<item id="R104" status="unpaid" expires="2022-12-31">
<type>... | xpath_write | stable | deterministic | q|xpath|20261022|4904 | 1,276 | 517 | 61d4c46b2f4d5e87f9642cff796f69e517a030f8558d98765da99b998f5381d0 | {"answer_form": "sentence", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpw:4", "gold": {"query": "//item[type='training']/../@name", "result": ["Fairhaven Logistics", "Thistle Group", "Stonebrook Labs"]}, "marker": "wrap:5", "presentation": "plain", "source": "XPath 1.0", "stratum": "XPath 1.0:parent_axis... |
queries-b6159885cb6d9cbb | <|start|><|user|>
Write an XPath 1.0 expression that returns the total of the fee values of the items in the group with code 'G1'.
<text>
<registry source="audit copy">
<group code="G1" name="Dunmore Supply">
<item id="R112" status="disputed">
<type>hardware</type>
<year>2022</year>
<fee>2050</... | xpath_write | stable | deterministic | q|xpath|20261022|8472 | 1,794 | 721 | 78a410f9688a7d6d36ec19e7a347e54eff9372287de8d197bdaf0932fa50f41d | {"answer_form": "sentence", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpw:0", "gold": {"query": "sum(//group[@code='G1']/item/fee)", "result": "4400"}, "marker": "wrap:4", "presentation": "chat", "source": "XPath 1.0", "stratum": "XPath 1.0:sum_group", "template": "sum_group"} |
queries-49a8fd8e402e5832 | What does jq 1.8 output (with -c, one JSON value per line) for this filter and input?
Text:
{
"meta": {
"source": "audit sample",
"generated": "2024-08-16"
},
"loans": [
{
"id": 101,
"purpose": "personal",
"year": 2018,
"principal": null,
"status": "paid_off"
},
... | jq_predict | stable | deterministic | q|jq|20261022|3019 | 1,890 | 650 | 84a691cdb1f60d2f36ca1da6f9a13683b9466fbcad0a28ae1be490d70d00e292 | {"answer_form": "structured", "data_form": "json", "dialect": "jq 1.8", "frame_id": "jqp:0", "gold": {"query": ".loans | map(select(.status == \"paid_off\")) | length", "result": [4]}, "marker": "wrap:2", "presentation": "plain", "source": "jq 1.8", "stratum": "jq 1.8:count_status", "template": "count_status"} |
queries-22f5b48f0ad96146 | <|start|><|user|>
What does jq 1.8 output (with -c, one JSON value per line) for this filter and input?
Excerpt:
{
"meta": {
"source": "registry export",
"generated": "2023-08-12"
},
"permits": [
{
"id": 102,
"permit_type": "export",
"year": 2016,
"fee": 850,
"note": "re... | jq_predict | stable | deterministic | q|jq|20261022|4881 | 1,924 | 754 | 14b548420e5ab063a12e02dbf3c26254b8a5d5d008da70d1ff1a60a5d2ce9c60 | {"answer_form": "sentence", "data_form": "json", "dialect": "jq 1.8", "frame_id": "jqp:0", "gold": {"query": "[.permits[] | select(.status == \"pending\") | {id, fee}]", "result": [[{"fee": 2250, "id": 113}, {"fee": 2150, "id": 124}]]}, "marker": "wrap:5", "presentation": "chat", "source": "jq 1.8", "stratum": "jq 1.8:... |
queries-ba8ef0cc238cfe53 | For the document below, what XPath 1.0 expression returns the name attributes of the groups containing an item whose type is 'supplies' (each group once, in document order)?
Excerpt:
<registry source="field office">
<group code="G1" name="Birchwood Holdings">
<item id="R102" status="under_review">
<type>re... | xpath_write | stable | deterministic | q|xpath|20261022|3866 | 1,385 | 549 | 1acfd6325352c5f337eb19587a6f96d8677965823f92460a500e90be82f9b7ce | {"answer_form": "sentence", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpw:5", "gold": {"query": "//item[type='supplies']/../@name", "result": ["Birchwood Holdings", "Thistle Labs"]}, "marker": "wrap:5", "presentation": "plain", "source": "XPath 1.0", "stratum": "XPath 1.0:parent_axis", "template": "paren... |
queries-4373bdc102a83be7 | Q: jq 1.8, compact mode: show the output of the filter below.
---
{
"meta": {
"source": "audit sample",
"generated": "2023-03-13"
},
"contracts": [
{
"id": 101,
"category": "research",
"year": 2018,
"value": 2500,
"status": "terminated",
"note": "refiled"
},
... | jq_predict | stable | deterministic | q|jq|20261022|7807 | 1,494 | 548 | 83aab56a48e1d6da4021f37dff7b17df23fd5626aef87a6dd778ffc5f6f1e235 | {"answer_form": "structured", "data_form": "json", "dialect": "jq 1.8", "frame_id": "jqp:4", "gold": {"query": ".contracts[] | select(.value > 1500) | .id", "result": [101, 104, 108, 110]}, "marker": "wrap:1|Q:", "presentation": "qa", "source": "jq 1.8", "stratum": "jq 1.8:select_gt", "template": "select_gt"} |
queries-4f91e4373ffd7aea | <|start|><|user|>
Predict the result of the XPath 1.0 query on the XML shown (node-sets as the list of string-values).
<text>
<registry source="annual export"><group code="G1" name="Eastgate Supply"><item id="R106" status="terminated" expires="2021-12-31"><type>research</type><year>2017</year><fee>1750</fee></item><it... | xpath_predict | stable | deterministic | q|xpath|20261022|6220 | 1,255 | 503 | fbbf9ad3e3986ea3a16453f173307d54a32224f0f62c9c5d52a14b5257f67bca | {"answer_form": "structured", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpp:2", "gold": {"query": "//item[@status='terminated']/type/text()", "result": ["research", "research", "consulting"]}, "marker": "wrap:4", "presentation": "chat", "source": "XPath 1.0", "stratum": "XPath 1.0:text_values", "template... |
queries-a6315b68fe0da37c | Run this SQLite 3 query in your head against the tables below. What does it return?
Passage:
CREATE TABLE vendors (id INTEGER PRIMARY KEY, name TEXT NOT NULL, tier TEXT);
CREATE TABLE contracts (id INTEGER PRIMARY KEY, vendor_id INTEGER REFERENCES vendors(id), category TEXT, year INTEGER, value INTEGER, status TEXT);
... | sql_predict | stable | deterministic | q|sql|20261022|3452 | 1,105 | 466 | fee24fffe5dce258d527cbb4460a6285d726e849d03036e8a1eff11b4b1e42ef | {"answer_form": "structured", "data_form": "ddl", "dialect": "SQLite 3", "frame_id": "sqlp:0", "gold": {"columns": ["id"], "result": [[101], [114], [116], [124]], "sql": "SELECT id FROM contracts WHERE value > (SELECT AVG(value) FROM contracts) ORDER BY id;"}, "marker": "wrap:3", "presentation": "plain", "source": "SQL... |
queries-2f69d154468d66fb | <|start|><|user|>
Predict the output of running jq -c with the filter below on this JSON (jq 1.8).
Passage:
{"meta": {"source": "batch 7", "generated": "2021-09-16"}, "filings": [{"id": 102, "court": "tax", "year": 2022, "fee": 2600, "status": "closed", "note": "refiled"}, {"id": 105, "court": "bankruptcy", "year": 20... | jq_predict | stable | deterministic | q|jq|20261022|9756 | 1,361 | 565 | add64ad909ce9965424f817d566a1f6ea67188e32b08e094215809e299bd470a | {"answer_form": "structured", "data_form": "json", "dialect": "jq 1.8", "frame_id": "jqp:1", "gold": {"query": ".filings | map(select(.fee != null)) | max_by(.fee) | .id", "result": [109]}, "marker": "wrap:3", "presentation": "chat", "source": "jq 1.8", "stratum": "jq 1.8:max_by", "template": "max_by"} |
queries-4efe7905a664ef31 | Query: Write SQL (SQLite 3) for this report request.
<text>
Table `customers`:
| id | name | region |
|---|---|---|
| 1 | Juniper Partners | north |
| 2 | Harlow Services | south |
| 3 | Zephyr Trust | east |
Table `invoices`:
| id | customer_id | product | year | amount | status |
|---|---|---|---|---|---|
| 100 |... | sql_write | stable | deterministic | q|sql|20261022|4941 | 709 | 302 | ca814d4a28b98cebb3ac77729aca203a2ec1fe6d4fac3f3b9d9bed148e976a8c | {"answer_form": "code", "data_form": "md", "dialect": "SQLite 3", "frame_id": "sqlw:6", "gold": {"result": [], "sql": "SELECT name FROM customers WHERE NOT EXISTS (SELECT 1 FROM invoices WHERE invoices.customer_id = customers.id) ORDER BY id;"}, "marker": "wrap:4|Query:", "presentation": "qa", "source": "SQLite 3", "st... |
queries-11e3f80abebb800f | Request: Give a jq filter (jq 1.8, run as jq -c) that will output the id of each record in .loans whose principal is greater than 2650, one per line.
---
{"meta": {"source": "audit sample", "generated": "2020-02-16"}, "loans": [{"id": 100, "purpose": "personal", "year": 2023, "principal": 1850, "status": "paid_off"}, ... | jq_write | stable | deterministic | q|jq|20261022|4189 | 1,197 | 481 | dc102190a1f1bd16a6dc6791a52f03bfbdb688c8388a0305fdd69317c41541ab | {"answer_form": "code", "data_form": "json", "dialect": "jq 1.8", "frame_id": "jqw:2", "gold": {"query": ".loans[] | select(.principal > 2650) | .id", "result": []}, "marker": "wrap:1|Request:", "presentation": "qa", "source": "jq 1.8", "stratum": "jq 1.8:select_gt", "template": "select_gt"} |
queries-ffc9fc32bd6f1c05 | Write a SQLite 3 query for the request below, using the tables shown.
Excerpt:
Table `customers`:
| id | name | region |
|---|---|---|
| 1 | Harlow Holdings | south |
| 2 | Cobalt Group | west |
| 3 | Harlow Trust | north |
| 4 | Kestrel Logistics | south |
Table `invoices`:
| id | customer_id | product | year | am... | sql_write | stable | deterministic | q|sql|20261022|69 | 925 | 384 | 5bf10c9c8bf0710200bfdf5c5898d38afdf8f7b89d83e9f7e2dea9b21ba4ab6e | {"answer_form": "code", "data_form": "md", "dialect": "SQLite 3", "frame_id": "sqlw:0", "gold": {"result": [[100, "Kestrel Logistics"], [104, "Kestrel Logistics"], [106, "Kestrel Logistics"], [112, "Harlow Holdings"], [116, "Kestrel Logistics"], [119, "Kestrel Logistics"]], "sql": "SELECT invoices.id, customers.name FR... |
queries-8b7f862bc735f243 | Predict the output of the query (SQLite 3) for the tables shown.
Passage:
CREATE TABLE customers (id INTEGER PRIMARY KEY, name TEXT NOT NULL, region TEXT);
CREATE TABLE invoices (id INTEGER PRIMARY KEY, customer_id INTEGER REFERENCES customers(id), product TEXT, year INTEGER, amount INTEGER, status TEXT);
INSERT INTO ... | sql_predict | stable | deterministic | q|sql|20261022|1219 | 949 | 435 | 70a3bd8e05cbe7b661e8a50c568397df1f27a55a6c7cc4657878548969fd9095 | {"answer_form": "structured", "data_form": "ddl", "dialect": "SQLite 3", "frame_id": "sqlp:2", "gold": {"columns": ["name", "n"], "result": [["Cobalt Labs", 2], ["Ridgeline Trust", 0], ["Larkspur Trust", 3], ["Ironside Works", 1]], "sql": "SELECT customers.name, COUNT(invoices.id) AS n FROM customers LEFT JOIN invoices... |
queries-3a421199d1f597a8 | <|start|><|user|>
Given the input JSON, what values does the jq filter produce? Use jq 1.8 semantics; one result per line.
Passage:
{
"meta": {
"source": "audit sample",
"generated": "2023-07-17"
},
"loans": [
{
"id": 100,
"purpose": "auto",
"year": 2021,
"principal": 2200,
... | jq_predict | stable | deterministic | q|jq|20261022|8983 | 1,951 | 687 | b6b76e4cdf9d127aac13d1ea6dd9551f324b943b0c62bc0e02dc61a1b9d15e07 | {"answer_form": "structured", "data_form": "json", "dialect": "jq 1.8", "frame_id": "jqp:3", "gold": {"query": ".loans[] | select(.principal > 1250) | .id", "result": [100, 104, 108, 109]}, "marker": "wrap:3", "presentation": "chat", "source": "jq 1.8", "stratum": "jq 1.8:select_gt", "template": "select_gt"} |
queries-429098f67ee7a986 | <|start|><|user|>
Evaluate the jq filter on the input. Print each output on its own line, as jq -c would.
```
{
"meta": {
"source": "batch 7",
"generated": "2020-03-19"
},
"loans": [
{
"id": 101,
"purpose": "mortgage",
"year": 2022,
"principal": 2750,
"status": "late",
... | jq_predict | stable | deterministic | q|jq|20261022|8774 | 1,877 | 674 | 43ee0a6e3d988926bf191ffa53450cd1bbb4f0b6f8646509b52a20276260c040 | {"answer_form": "sentence", "data_form": "json", "dialect": "jq 1.8", "frame_id": "jqp:5", "gold": {"query": "[.loans[] | select(has(\"note\")) | .id]", "result": [[101, 108, 114, 120]]}, "marker": "wrap:6", "presentation": "chat", "source": "jq 1.8", "stratum": "jq 1.8:has_note", "template": "has_note"} |
queries-07be0ba073b2edc0 | Query: Apply the XPath 1.0 expression to the document and give the result.
---
<registry source="field office">
<group code="G1" name="Upland Associates">
<item id="R102" status="open">
<type>district</type>
<year>2022</year>
<fee>2250</fee>
</item>
<item id="R128" status="closed" expir... | xpath_predict | stable | deterministic | q|xpath|20261022|6822 | 1,621 | 663 | f2730e4754d45e9e74356e5afb9067fa7a547679aa4e05d06bc204a6e1b9132a | {"answer_form": "structured", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpp:3", "gold": {"query": "//item[type='district']/../@name", "result": ["Upland Associates", "Thistle Labs", "Yarrow Works", "Zephyr Group"]}, "marker": "wrap:1|Query:", "presentation": "qa", "source": "XPath 1.0", "stratum": "XPath... |
queries-140f77c87ee68913 | Question: Give the XPath 1.0 query that returns the name attributes of the groups containing an item whose type is 'software' (each group once, in document order).
Text:
<registry source="audit copy">
<group code="G1" name="Thistle Works">
<item id="R107" status="void">
<type>software</type>
<year>20... | xpath_write | stable | deterministic | q|xpath|20261022|8187 | 1,710 | 674 | 62b64fc3161fac269e095736ec40430672b8f54904e8efd6b31b9f97c459a61a | {"answer_form": "code", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpw:2", "gold": {"query": "//item[type='software']/../@name", "result": ["Thistle Works", "Upland Supply", "Ironside Logistics"]}, "marker": "wrap:2|Question:", "presentation": "qa", "source": "XPath 1.0", "stratum": "XPath 1.0:parent_axis... |
queries-010fc51f303c7dc1 | Question: Write SQL (SQLite 3) for this report request.
Text:
CREATE TABLE lenders (id INTEGER PRIMARY KEY, name TEXT NOT NULL, charter TEXT);
CREATE TABLE loans (id INTEGER PRIMARY KEY, lender_id INTEGER REFERENCES lenders(id), purpose TEXT, year INTEGER, principal INTEGER, status TEXT);
INSERT INTO lenders VALUES (1... | sql_write | stable | deterministic | q|sql|20261022|887 | 1,539 | 693 | 898d99abe660ab82624b15bdd2509f78eb382cfee88cdab8c1d1f6543f960b68 | {"answer_form": "sentence", "data_form": "ddl", "dialect": "SQLite 3", "frame_id": "sqlw:6", "gold": {"result": [[700, 2000, 1362.5]], "sql": "SELECT MIN(principal) AS lo, MAX(principal) AS hi, ROUND(AVG(principal), 2) AS mean FROM loans WHERE status = 'current';"}, "marker": "wrap:2|Question:", "presentation": "qa", "... |
queries-38cc61057d609fc8 | <|start|><|system|>
You check your work before answering.<|eot|>
<|user|>
Write SQL (SQLite 3) for this report request.
---
Table `customers`:
| id | name | region |
|---|---|---|
| 1 | Zephyr Trust | south |
| 2 | Vantage Works | east |
| 3 | Zephyr Supply | north |
| 4 | Birchwood Group | east |
| 5 | Kestrel Group... | sql_write | stable | deterministic | q|sql|20261022|6554 | 1,103 | 454 | 136bce49336b5ffc8f01c08025b8be8c2a44847b7ed478517ef5fa3cecc1972b | {"answer_form": "sentence", "data_form": "md", "dialect": "SQLite 3", "frame_id": "sqlw:6", "gold": {"result": [["hardware", 4, 5450], ["software", 5, 5650], ["support", 1, 200]], "sql": "SELECT product, COUNT(*) AS n, SUM(amount) AS total FROM invoices GROUP BY product ORDER BY product;"}, "marker": "wrap:1", "present... |
queries-2f2120710e3b7683 | Task: Predict the output of the query (SQLite 3) for the tables shown.
Excerpt:
CREATE TABLE customers (id INTEGER PRIMARY KEY, name TEXT NOT NULL, region TEXT);
CREATE TABLE invoices (id INTEGER PRIMARY KEY, customer_id INTEGER REFERENCES customers(id), product TEXT, year INTEGER, amount INTEGER, status TEXT);
INSERT... | sql_predict | stable | deterministic | q|sql|20261022|8166 | 1,138 | 513 | ff41c35e8a6b0f7da1f70a03b7fc4f8177f3d7171da490dc2ff7e6251bac55b6 | {"answer_form": "structured", "data_form": "ddl", "dialect": "SQLite 3", "frame_id": "sqlp:2", "gold": {"columns": ["name", "n"], "result": [["Juniper Partners", 2], ["Ridgeline Labs", 3], ["Ironside Labs", 1], ["Northfield Logistics", 2], ["Yarrow Partners", 1]], "sql": "SELECT customers.name, COUNT(invoices.id) AS n ... |
queries-d1c4d72ad3606ecf | Query: Run this SQLite 3 query in your head against the tables below. What does it return?
```
Table `parties`:
| id | name | kind |
|---|---|---|
| 1 | Eastgate Group | individual |
| 2 | Yarrow Partners | agency |
| 3 | Cobalt Works | individual |
| 4 | Dunmore Labs | agency |
| 5 | Juniper Associates | individual ... | sql_predict | stable | deterministic | q|sql|20261022|801 | 902 | 392 | 94cc8bc0643aee758ab2363f654677d19a37996a049624761026a781df23550e | {"answer_form": "structured", "data_form": "md", "dialect": "SQLite 3", "frame_id": "sqlp:0", "gold": {"columns": ["id", "band"], "result": [[100, "unknown"], [104, "high"], [108, "low"], [109, "high"], [112, "high"], [116, "unknown"], [120, "unknown"]], "sql": "SELECT id, CASE WHEN fee IS NULL THEN 'unknown' WHEN fee ... |
queries-4f4377868c7dc82f | Instruction: Predict the result of the JSONPath query (RFC 9535) on this JSON: give the values as an array, in order.
Text:
{
"meta": {
"source": "registry export",
"generated": "2022-04-16"
},
"permits": [
{
"id": 101,
"permit_type": "fishing",
"year": 2019,
"fee": 1650,
... | jsonpath_predict | stable | deterministic | q|jsonpath|20261022|1828 | 1,421 | 511 | dfd844a98c01f88f3b09b84b213fab09adbb3984cf8b0c391887e3e018213235 | {"answer_form": "structured", "data_form": "json", "dialect": "JSONPath (RFC 9535)", "frame_id": "jsonpathp:2", "gold": {"query": "$.permits[-1].id", "result": [115]}, "marker": "wrap:2|Instruction:", "presentation": "qa", "source": "JSONPath (RFC 9535)", "stratum": "JSONPath (RFC 9535):neg_index", "template": "neg_ind... |
queries-b53e568f4f1353b1 | <|start|><|user|>
Apply the XPath 1.0 expression to the document and give the result.
Excerpt:
<registry source="annual export"><group code="G1" name="Juniper Works"><item id="R107" status="expired"><type>fishing</type><year>2021</year><fee>2100</fee></item><item id="R126" status="active"><type>fishing</type><year>201... | xpath_predict | stable | deterministic | q|xpath|20261022|565 | 1,408 | 586 | b8331c0c197f1f3c31f653faa81554d20c96a268becc5156d0edb3e79ff4065a | {"answer_form": "bare", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpp:3", "gold": {"query": "sum(//group[@code='G1']/item/fee)", "result": "4500"}, "marker": "wrap:5", "presentation": "chat", "source": "XPath 1.0", "stratum": "XPath 1.0:sum_group", "template": "sum_group"} |
queries-420b2027d06b52c4 | <|start|><|user|>
Give the RFC 9535 JSONPath expression to select the id of every element of the loans array.
<text>
{"meta": {"source": "registry export", "generated": "2022-05-13"}, "loans": [{"id": 101, "purpose": "personal", "year": 2016, "principal": 1550, "status": "paid_off", "note": "priority"}, {"id": 103, "p... | jsonpath_write | stable | deterministic | q|jsonpath|20261022|4368 | 1,373 | 517 | 0e3105ffe9f5fd75fda123590f14dbe88dc4c56c049e7b5512a60453478d94bc | {"answer_form": "code", "data_form": "json", "dialect": "JSONPath (RFC 9535)", "frame_id": "jsonpathw:2", "gold": {"query": "$.loans[*].id", "result": [101, 103, 106, 110, 114, 117, 119, 121, 124]}, "marker": "wrap:4", "presentation": "chat", "source": "JSONPath (RFC 9535)", "stratum": "JSONPath (RFC 9535):wild", "temp... |
queries-9919279d51af9862 | Request: Run the XPath 1.0 expression below on the XML and report its result.
Text:
<registry source="field office">
<group code="G1" name="Harlow Trust">
<item id="R102" status="void" expires="2022-12-31">
<type>hosting</type>
<year>2017</year>
<fee>850</fee>
</item>
<item id="R104" st... | xpath_predict | stable | deterministic | q|xpath|20261022|791 | 1,611 | 687 | ee641a8b14eb5afeeeef897591d10d0256a79f006c60268617983727177524ab | {"answer_form": "structured", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpp:5", "gold": {"query": "//item[not(@expires)]/@id", "result": ["R112", "R123", "R118", "R127", "R106", "R116"]}, "marker": "wrap:2|Request:", "presentation": "qa", "source": "XPath 1.0", "stratum": "XPath 1.0:not_attr", "template"... |
queries-2e599b8fa4f2376b | <|start|><|user|>
Give the XPath 1.0 query that returns the id attributes of items that have no expires attribute.
Text:
<registry source="audit copy">
<group code="G1" name="Juniper Trust">
<item id="R104" status="dismissed">
<type>patent</type>
<year>2023</year>
<fee>1800</fee>
</item>
... | xpath_write | stable | deterministic | q|xpath|20261022|3943 | 1,831 | 714 | 0384618d0e7ad0c1e105a72cc6163ca2d7e0df23172c1ebd9e1ad86bbc6d1abf | {"answer_form": "code", "data_form": "xml", "dialect": "XPath 1.0", "frame_id": "xpw:2", "gold": {"query": "//item[not(@expires)]/@id", "result": ["R104", "R112", "R126", "R127", "R131", "R109", "R117", "R101", "R120"]}, "marker": "wrap:2", "presentation": "chat", "source": "XPath 1.0", "stratum": "XPath 1.0:not_attr",... |
End of preview. Expand in Data Studio
ALEA query languages
Write a query or predict its output in SQLite 3, jq 1.8, JSONPath (RFC 9535 subset) or XPath 1.0 over small shown datasets. SQL items cover NULL logic, COUNT(*) versus COUNT(col), GROUP BY / HAVING, joins, DISTINCT, top-k with tie-breaks, subqueries, CASE and the NOT IN with NULL trap.
33,254 documents, 17,807,318 tokens (alea-v2b-sw32k, the ALEA R4 tokenizer, counting each document's start and end tokens).
Part of the ALEA R4 training mix
This dataset contains exactly the documents of this family that ALEA's R4 model trained on in its stable pretraining mixture (second data stage, one pass each). Every row is byte-identical to the trained text; text_sha256 is the SHA-256 of text.
Held-out evaluation items generated by the same code are not part of this release.
Composition
| task | documents | tokens |
|---|---|---|
sql_predict |
7,442 | 3,390,499 |
sql_write |
6,986 | 3,163,645 |
xpath_write |
3,550 | 2,046,115 |
xpath_predict |
3,481 | 2,015,852 |
jq_predict |
3,450 | 2,140,820 |
jq_write |
3,308 | 2,052,024 |
jsonpath_write |
2,531 | 1,490,030 |
jsonpath_predict |
2,506 | 1,508,333 |
11,651 documents use ALEA's chat template.
How it was made
Written by deterministic code from typed data with invented names and values; no language model was used. Gold is produced by executing the query in the real engine.
Sources
- Invented tables and documents
Verification
- Each template carries an independent Python reference implementation of its English description; gold is kept only when engine and reference agree on the shown data and on two perturbed copies.
- Decontamination: training rows that shared a word 13-gram with a held-out evaluation item of this or any sibling family were removed, and rows overlapping ALEA's evaluation sets were removed or logged according to ALEA's decontamination policy.
- Sampled records were reviewed blind by a separate reviewer who recomputed or re-derived the answers; defect patterns found in review were fixed in the generator, or the affected rows dropped, before the final build.
Columns
| column | content |
|---|---|
id |
row id from the generator (not always unique: a few families reuse an id across variants, and a handful of exact duplicate documents were trained twice and appear twice) |
text |
the document exactly as trained; chat documents use ALEA's chat template, whose control tokens appear as text (`< |
task |
task kind |
phase |
stable (trained in R4's stable pretraining mixture) |
generator |
what wrote the row (deterministic = code only) |
group_id |
source passage / record group; all variants of one group share it |
prompt_chars |
length in characters of the request part of text (0 for plain documents; null where not recorded) |
n_tokens |
tokens under alea-v2b-sw32k, including the document start and end tokens |
text_sha256 |
SHA-256 of text |
meta |
JSON: the generator's structured fields (source, presentation, answer form, gold values, checks) |
Known limitations
- The template pool is narrow; query shapes recur.
- The documents are generated and more repetitive than natural text.
License
CC-BY-4.0 for this compilation and the generated text. Public-domain and public-record source text keeps its status; Wikidata is CC0; OpenGloss and SHELF are CC-BY-4.0.
Published by the ALEA Institute.
- Downloads last month
- 15