alibaba/spring-ai-alibaba · error · IllegalArgumentException

评测集列配置错误，必须包含input和reference_output两列

Error message

评测集列配置错误，必须包含input和reference_output两列

What it means

Thrown by DatasetServiceImpl.create() when columnsConfig is missing or does not contain the required 'input' and 'reference_output' columns, as checked by hasRequiredColumns(). Evaluation datasets must have this fixed schema.

Solutions

1. Ensure columnsConfig includes both an 'input' column and a 'reference_output' column
2. Use the schema produced by the platform's dataset-creation UI/export as a template
3. Add client-side validation of column names before submitting
4. Call hasRequiredColumns-equivalent check (or the validation API) before create

Example fix

// before
columnsConfig = [{"name":"question"},{"name":"answer"}];
// after
columnsConfig = [{"name":"input"},{"name":"reference_output"}];

Set<String> names = columns.stream().map(ColumnDef::getName).collect(Collectors.toSet());
if (!names.contains("input") || !names.contains("reference_output")) { throw new IllegalArgumentException("columnsConfig must define input and reference_output"); }

boolean hasRequiredColumns(List<ColumnConfig> cols) { return cols != null && cols.stream().map(ColumnConfig::getName).collect(Collectors.toSet()).containsAll(Set.of("input","reference_output")); }

try { datasetService.create(req); } catch (IllegalArgumentException e) { if (e.getMessage().contains("列配置")) { showSchemaErrorToUser(); } else { throw e; } }

When it happens

Trigger: Creating a dataset whose columnsConfig JSON lacks 'input' or 'reference_output' column definitions; passing null columnsConfig; importing a dataset exported from another tool with a different schema.

Common situations: Hand-written columns config missing a column; renamed columns; upstream export format changed; deserialization typo in the columns key.

Understand the failure class

Background: Schema validation failed / invalid input schema: payload rejected because its shape doesn't match the expected schema — this error's family across 28 libraries.

Related errors

• 数据集版本数据量为0或不存在
• AgentCard or AgentCardProvider must be provided
• AgentScope Model must be provided for AgentScope routing…
• AgentScope routing flow requires at least one sub-agent
• appName cannot be null or empty

AI-assisted analysis of alibaba/spring-ai-alibaba@f82da0b50f (2026-09-09). Data as JSON: /api/errors/96576f2530e5c36a. Report an issue: GitHub.

Appendix: source

Thrown at spring-ai-alibaba-admin/spring-ai-alibaba-admin-server-start/src/main/java/com/alibaba/cloud/ai/studio/admin/service/impl/DatasetServiceImpl.java:54

    private static final String INPUT_COLUMN_TYPE = "input";
    private static final String REFERENCE_OUTPUT_COLUMN_TYPE = "reference_output";


    @Override
    @Transactional
    public Dataset create(DatasetCreateRequest request) {
        log.info("创建评测集: {}", request);

        if (!StringUtils.hasText(request.getName())) {
            throw new IllegalArgumentException("评测集名称不能为空");
        }


        if (request == null || request.getColumnsConfig() == null ||
                !hasRequiredColumns(request.getColumnsConfig())) {
            throw new IllegalArgumentException("评测集列配置错误，必须包含input和reference_output两列");
        }

        DatasetDO datasetDO = DatasetDO.builder()
                .name(request.getName())
                .description(request.getDescription())
                .columnsConfig(JSONObject.toJSONString(request.getColumnsConfig()))
                .build();

        datasetMapper.insert(datasetDO);
        log.info("评测集创建成功: {}", datasetDO);
        return Dataset.fromDO(datasetDO);
    }

    @Override
    public PageResult<Dataset> list(DatasetListRequest request) {
        log.info("查询评测集列表: {}", request);

        // 计算分页参数

View on GitHub (pinned to f82da0b50f)