Dataset Viewer
Auto-converted to Parquet Duplicate
problem
stringclasses
2 values
gtruth_answer
stringclasses
2 values
message_thinking
stringclasses
2 values
message_content
stringclasses
2 values
Given the following MYSQL schema: ```sql CREATE TABLE invoices ( invoice_id INT AUTO_INCREMENT PRIMARY KEY, customer_id INT NOT NULL, issued_date DATE NOT NULL, due_date DATE NOT NULL, amount DECIMAL(12,2) NOT NULL, paid_at DATETIME ); CREATE TABLE customers ( customer_id INT AUTO_INCREMEN...
```sql SELECT i.invoice_id, c.company_name, c.contact_email, i.issued_date, i.due_date, i.amount, DATEDIFF(CURDATE(), i.due_date) AS days_overdue FROM invoices i JOIN customers c ON c.customer_id = i.customer_id WHERE i.paid_at IS NULL AND i.due_date < CURDATE() ORDER BY days_overdue DESC;...
The user wants a MySQL query to find overdue unpaid invoices with customer name and days overdue. Let me write this query. Overdue unpaid invoices: due_date < current date, and paid_at IS NULL. Need customer company_name from customers table. Days overdue: DATEDIFF(current_date, due_date). I'll use CURDATE() for curr...
```sql SELECT i.invoice_id, c.company_name, i.due_date, i.amount, DATEDIFF(CURDATE(), i.due_date) AS days_overdue FROM invoices i JOIN customers c ON i.customer_id = c.customer_id WHERE i.due_date < CURDATE() AND i.paid_at IS NULL; ``` \boxed{Query above}
Given the following BIGQUERY schema: ```sql -- BigQuery table: `events.raw_events` -- Fields: -- event_id STRING -- user_id STRING -- event_name STRING -- event_time TIMESTAMP -- properties JSON STRING (e.g. '{"tags": ["premium", "trial"], "plan": "pro"}') -- tags ARRAY<STRING> ``` Write a BigQuery SQL qu...
```sql SELECT tag, COUNT(*) AS event_count FROM `events.raw_events`, UNNEST(tags) AS tag WHERE event_time >= TIMESTAMP_SUB(CURRENT_TIMESTAMP(), INTERVAL 30 DAY) GROUP BY tag ORDER BY event_count DESC; ``` **How it works:** BigQuery's `UNNEST(array_column) AS alias` in the `FROM` clause (cross join unnest) exp...
The user wants a BigQuery SQL query to count events per tag for the last 30 days. The tags field is ARRAY<STRING>. Need to unnest the array and count. Key points: - Table: `events.raw_events` - Fields: event_id, user_id, event_name, event_time, properties (JSON STRING), tags (ARRAY<STRING>) - Need last 30 days relativ...
\boxed{WITH filtered AS (SELECT event_id, tags FROM \`events.raw_events\` WHERE event_time >= TIMESTAMP_SUB((SELECT MAX(event_time) FROM \`events.raw_events\`), INTERVAL 30 DAY)) SELECT tag, COUNT(*) AS event_count FROM filtered, UNNEST(tags) AS tag GROUP BY tag ORDER BY event_count DESC;}

SQL-SFT-V1

Dataset distilasi tentang pelaporan pajak di Indonesia dalam Bahasa Indonesia, dibuat untuk belajar membangun LLM sendiri. Formatnya mengikuti rasbt/math_distill.

Construction

  • Soal disediakan sendiri oleh pemilik dataset (file problems.jsonl) beserta kunci jawaban singkat (gtruth_answer).
  • Jawaban (thinking + jawaban akhir) dihasilkan oleh model teacher lewat router API; verifikasi otomatis terhadap kunci jawaban: VERIFY=True, ONLY_CORRECT=True, REQUIRE_THINKING=True.

Akurasi teacher sebelum filter:

Model teacher Sampel Benar Dengan thinking
nvidia/nemotron-3-ultra-550b-a55b 24 3 (12.5%) 24

Files

Fields

Each example contains:

  • problem: soal perpajakan
  • gtruth_answer: jawaban singkat (kunci)
  • message_thinking: thinking trace dari teacher
  • message_content: jawaban akhir dari teacher

Keduanya dapat digabung, sama seperti math_distill:

complete_answer = f"<think>{data['message_thinking']}</think>\n\n{data['message_content']}"

Dataset Statistics

Dihitung dengan scripts/average_field_lengths_json.py, misalnya:

python scripts/average_field_lengths_json.py --json_path data/pajak-eval.json --tokenizer Qwen/Qwen3-0.6B

data/pajak-eval.json

Records: 1
Tokenizer: Qwen/Qwen3-0.6B
Field             AvgTokens  MinTokens  MaxToken  Count
gtruth_answer        240.00        240       240      1
message_content      231.00        231       231      1
message_thinking     624.00        624       624      1
problem              112.00        112       112      1

Verified-correct answers: 1/1 (100.0%)

data/pajak-train.json

Records: 2
Tokenizer: Qwen/Qwen3-0.6B
Field             AvgTokens  MinTokens  MaxToken  Count
gtruth_answer        146.00        134       158      2
message_content       78.50         72        85      2
message_thinking     556.00        155       957      2
problem              123.00        115       131      2

Verified-correct answers: 2/2 (100.0%)

data/pajak-train_4000.json

Records: 2
Tokenizer: Qwen/Qwen3-0.6B
Field             AvgTokens  MinTokens  MaxToken  Count
gtruth_answer        146.00        134       158      2
message_content       78.50         72        85      2
message_thinking     556.00        155       957      2
problem              123.00        115       131      2

Verified-correct answers: 2/2 (100.0%)

data/nvidia-nemotron-3-ultra-550b-a55b-pajak-train.json

Records: 2
Tokenizer: Qwen/Qwen3-0.6B
Field             AvgTokens  MinTokens  MaxToken  Count
gtruth_answer        146.00        134       158      2
message_content       78.50         72        85      2
message_thinking     556.00        155       957      2
problem              123.00        115       131      2

Verified-correct answers: 2/2 (100.0%)

Intended use

Untuk pembelajaran pribadi: melatih/distilasi LLM kecil.

Usage

from datasets import load_dataset

repo_id = "bloksbgt/SQL-SFT-V1"

train = load_dataset(repo_id, "pajak_train", split="train")
eval_set = load_dataset(repo_id, "pajak_eval", split="train")
train_4000 = load_dataset(repo_id, "pajak_train_4000", split="train")

print(len(train), len(eval_set), len(train_4000))
print(train[0].keys())
Downloads last month
16