Improving SQL Generation with Schema-Aware Examples

Develop the training and evaluation logic behind qwen32b-analytics-sql. The reliable path is to make inputs, assumptions, and acceptance checks visible before producing a new adapter or deployment artifact. This note uses an illustrative workflow rather than a claim about a particular organization or production outcome.

Establish the boundary

Define a compact schema context format with tables, columns, keys, and dialect. Start with a versioned source set and preserve stable identifiers through every transformation. That makes it possible to compare a later output with the exact examples, configuration, and review rule that produced the earlier one. Record deterministic failures separately from reviewer judgment so that a retry does not silently alter the work.

Use a small holdout slice for this check. It should include ordinary examples, rare cases, and deliberately awkward inputs that reveal boundary conditions. Keep the result next to the source digest and configuration, then promote only the change that improves the intended condition without weakening the known constraints.

Pair questions with read-only SQL and explicit assumptions

Pair questions with read-only SQL and explicit assumptions. Start with a versioned source set and preserve stable identifiers through every transformation. That makes it possible to compare a later output with the exact examples, configuration, and review rule that produced the earlier one. Record deterministic failures separately from reviewer judgment so that a retry does not silently alter the work.

Include ambiguous questions that require clarification

Include ambiguous questions that require clarification. Start with a versioned source set and preserve stable identifiers through every transformation. That makes it possible to compare a later output with the exact examples, configuration, and review rule that produced the earlier one. Record deterministic failures separately from reviewer judgment so that a retry does not silently alter the work.

Parse SQL for forbidden write operations before execution

Parse SQL for forbidden write operations before execution. Start with a versioned source set and preserve stable identifiers through every transformation. That makes it possible to compare a later output with the exact examples, configuration, and review rule that produced the earlier one. Record deterministic failures separately from reviewer judgment so that a retry does not silently alter the work.

Evaluate syntax, referenced columns, execution, and result shape on a disposable database

Evaluate syntax, referenced columns, execution, and result shape on a disposable database. Start with a versioned source set and preserve stable identifiers through every transformation. That makes it possible to compare a later output with the exact examples, configuration, and review rule that produced the earlier one. Record deterministic failures separately from reviewer judgment so that a retry does not silently alter the work.

Why execution success alone is not semantic correctness

Explain why execution success alone is not semantic correctness. Start with a versioned source set and preserve stable identifiers through every transformation. That makes it possible to compare a later output with the exact examples, configuration, and review rule that produced the earlier one. Record deterministic failures separately from reviewer judgment so that a retry does not silently alter the work.

Practical check

SQLite setup and Python safety checks using standard libraries. The following minimal command or script is intended as a preflight check. It is deliberately small: its job is to expose missing structure before a longer run or a release review.

import json
import sys

for number, line in enumerate(sys.stdin, 1):
    record = json.loads(line)
    if not isinstance(record, dict):
        raise SystemExit(f"line {number}: expected object")
    print(json.dumps(record, ensure_ascii=False))

Release decision

A useful release record names the input version, the transformation or runtime configuration, the validation slice, and the remaining limitation. Executing generated SQL against production or accepting write statements. When evidence is incomplete, keep the artifact private, add the missing check, and preserve the failing example as a future regression case.

Keep the workflow idempotent. Repeating the same preparation step on the same input should not append metadata, discard additional detail, or change stable identifiers. Idempotence makes retries safe and turns unexpected differences into visible defects.

Review results by failure category instead of relying on a single blended score. A small number of high-risk failures can matter more than an improvement on common cases, especially when the adapter feeds a downstream queue or release process.

The final comparison should keep decoding, prompt formatting, and file versions fixed. If several variables move together, the result may look better while leaving the actual cause of the improvement unknown.

Retain both accepted and rejected examples. A concise rejected case with an expected outcome is a durable test asset and prevents the next contributor from rediscovering the same edge condition.