Data answer gaps: six recorded examples
The downloads contain six original inputs, their expected labels and one recorded API observation per input revision. Disagreements stay visible.
Provide the question, catalog notes and query report. Distinguish a missing field, ambiguous request and empty row list from a valid answer such as a count of zero.
The input is a report about a data question, not a live database connection. A missing field may require a new data source, an ambiguous request needs a clearer definition, and a completed empty list may simply mean the filters matched nothing.
The result must be read in the context of the question. A count query returning zero has supplied an answer. This check does not verify that the SQL was correct or that the database contains every real-world record; the stated catalog and query status are evidence supplied by the caller.
This original report-reading exercise is inspired by the cited CSV-ontology video. The video shows interface messages, not verified database behavior. These synthetic inputs do not reproduce its implementation, authenticate its results or connect to a database.

Use the sample to understand the rule, then test your own input for data answer gaps.

Provide the question, catalog notes and query report.

Open the decision rules and keep the intended distinction between Missing field, No matches, Ambiguous request, Available, Unclear. Edit the choices only when testing your own rule.

Use a sample first, then try your own input. Compare the selected label with the other returned probabilities.

Review Missing field, No matches, Ambiguous request before using a result. The label is a decision about the supplied information, not an executed action.
| Original input | Observed label | Leading probability |
|---|---|---|
| The required fact is not stored | Missing field | 100% |
| A supported query returns no vehicles | No matches | 98.0% |
| Best has no agreed measure | Ambiguous request | 100% |
| An answer is present | Available | 100% |
| Zero is the requested count | Available | 95.0% |
| Execution outcome is unknown | Unclear | 88.0% |
These are six original synthetic inputs, tested once per request revision. The percentage is the leading option probability, not measured accuracy. Authored expectations and any disagreements remain in the downloadable evidence. 6 of 6 results matched the authored fixture expectations.Verified
The downloads contain six original inputs, their expected labels and one recorded API observation per input revision. Disagreements stay visible.
The single-request SDK runner uses your private TYPESAFE_API_KEY. The offline consumer reads saved observations, checks request hashes and prints a pending output. It does not call the model or change any external system.
# Offline: inspect the downloaded saved observations. No API key is needed.
python preview_choice_workflow.py samples.json saved-api-results.json
# Optional live request: install the SDK and set TYPESAFE_API_KEY privately.
npm install @typesafe-ai/sdk@0.6.0
node run-choice-sample.mjs samples.json case-1
# The live response prints to the terminal; it does not update saved-api-results.json.The input contract and these limits define what the result means.
It means the supplied completed row-list query returned no qualifying records. The underlying fields may be available and other records may exist. Do not describe that outcome as a missing column or unavailable metric.
The question asks how many, and the report supplies that count. A zero value is different from having no report, an unavailable field or a request whose meaning has not been defined.
No. It shows interface text describing a known data gap and a separate no-matching-records outcome. These original examples explain that distinction without verifying the source application or its data.
It describes the distribution over the supplied choices for this request. These six synthetic examples are functional tests with authored expectations, not a production accuracy benchmark.
No. The demo displays a label and distribution. The downloadable consumer produces a pending preview packet. Application code and a separate authorization policy would control any real action.
Edit the example above to see how a changed fact affects the decision, or open the Playground for another task.