weaviate/weaviate · error

aggregate sparse search: %w

Error message

aggregate sparse search: %w

What it means

Wrapper in the aggregator's hybrid() closure: when the sparse (BM25) leg of hybrid search, executed via bm25Objects over the inverted index, fails, the error is re-wrapped as 'aggregate sparse search: %w'.

Source

Thrown at adapters/repos/db/aggregator/grouper.go:137

	} else {
		ids = allowList.Slice()
	}

	return ids, err
}

func (g *grouper) hybrid(ctx context.Context, allowList helpers.AllowList, modules *modules.Provider) ([]uint64, error) {
	g.setDefaultObjectLimit()

	sparseSearch := func() ([]*storobj.Object, []float32, error) {
		kw, err := g.buildHybridKeywordRanking()
		if err != nil {
			return nil, nil, fmt.Errorf("build hybrid keyword ranking: %w", err)
		}

		sparse, dists, err := g.bm25Objects(ctx, kw, allowList)
		if err != nil {
			return nil, nil, fmt.Errorf("aggregate sparse search: %w", err)
		}

		return sparse, dists, nil
	}

	denseSearch := func(vec models.Vector) ([]*storobj.Object, []float32, error) {
		res, dists, err := g.objectVectorSearch(ctx, vec, allowList)
		if err != nil {
			return nil, nil, fmt.Errorf("aggregate grouped dense search: %w", err)
		}

		return res, dists, nil
	}

	res, err := hybrid.Search(ctx, &hybrid.Params{
		HybridSearch: g.params.Hybrid,
		Keyword:      nil,
		Class:        g.params.ClassName.String(),

View on GitHub (pinned to 75aa4b6d11)

Solutions

1. Read the wrapped cause for the underlying BM25 failure.
2. Verify shard health and run recovery/repair on the affected collection if index errors are logged.
3. Check stopword configuration and tokenization settings in InvertedIndexConfig.
4. Retry the aggregation; if persistent, reindex or restore the affected shard from backup.

// ensure collection has text/searchable properties and healthy shards before hybrid aggregation
if len(class.Properties) == 0 { skip sparse leg }

err := runHybridAggregate(); if errors.Is/contains inverted-index or shard errors { retry with backoff; then trigger shard repair }

When it happens

Trigger: Hybrid aggregation where BM25F search over the collection's inverted index returns an error — inverted index corruption, missing shard, tokenizer/stopword configuration problems, or class lookup failure inside bm25Objects.

Common situations: Damaged or compacting LSM inverted indexes, searching a class whose searchable properties configuration changed, or shard unavailability on a node.

Related errors

• build hybrid keyword ranking: %w
• aggregate sparse search: %w
• bm25 objects: %w
• No properties provided, and no indexed properties found in c
• failed to extract hybrid params: %w

AI-assisted analysis of weaviate/weaviate@75aa4b6d11 (2026-09-04). Data as JSON: /api/errors/39986296891cf33a. Report an issue: GitHub.