dgraph-io/dgraph · error

NQuad failed sanity check. Subject

Error message

NQuad failed sanity check. Subject: %q, Predicate: %q, ObjectId: %q

What it means

ParseRDF validates each parsed NQuad before returning it. If the subject, predicate, or object ID contains characters that the sane() check rejects (control/whitespace/special characters), the NQuad cannot be safely indexed and the parse fails. The full input line is deliberately omitted from the message because RDF lines may contain sensitive data.

Solutions

1. Inspect the Subject/Predicate/ObjectId values quoted in the error and remove or escape invalid characters (spaces, control chars).
2. Ensure IRIs are wrapped in angle brackets <...> and blank nodes use the _:name form in the input line.
3. Pre-validate NQuads with the sane() check (or api.ValidateNQuad) before feeding lines to ParseRDF.
4. If the line is machine-generated, fix the upstream generator to emit canonical N-Quads.

Example fix

// before
ParseRDF("<http://ex/s> has name Alice .")            // invalid predicate 'has name'
// after
ParseRDF("<http://ex/s> <http://ex/name> \"Alice\" .")  // bracketed predicate, literal object

func validNQuadLine(line string) bool {
	for _, part := range []string{subject, predicate, objectId} {
		if !sane(part) { return false }
	}
	return true
}

func isSane(s string) bool { return sane(s) }

rnq, err := ParseRDF(line)
if err != nil {
	if strings.Contains(err.Error(), "failed sanity check") {
		log.Warn("skipping invalid NQuad", "err", err)
		return nil // skip or quarantine line
	}
	return err
}

When it happens

Trigger: Calling ParseRDF (or Parse/ParseRDFs) with an N-Quads line whose subject, predicate, or object ID contains invalid characters such as spaces, control characters, or unescaped special chars. Also occurs when a subject lacks a valid <...> or _:name form.

Common situations: Importing RDF dumps exported from other triple stores with loosely validated identifiers; hand-written .nq files with typos like unbracketed IRIs or stray whitespace; programmatic NQuad construction that skips validation before serialization.

Related errors

• empty facetKeys not allowed
• empty variable name in function call
• expected
• expected
• expected ')', found

AI-assisted analysis of dgraph-io/dgraph@759e242be6 (2026-09-01). Data as JSON: /api/errors/d6b3cea7820846d9. Report an issue: GitHub.

Appendix: source

Thrown at chunker/rdf_parser.go:225

		return rnq, fmt.Errorf("invalid end of input. Input: [%s]", line)
	}
	if isCommentLine {
		return rnq, ErrEmpty
	}
	// We only want to set default value if we have seen ObjectValue within "" and if we didn't
	// already set it.
	if seenOval && rnq.ObjectValue == nil {
		rnq.ObjectValue = &api.Value{Val: &api.Value_DefaultVal{DefaultVal: oval}}
	}
	if len(rnq.Subject) == 0 || len(rnq.Predicate) == 0 {
		return rnq, fmt.Errorf("empty required fields in NQuad. Input: [%s]", line)
	}
	if len(rnq.ObjectId) == 0 && rnq.ObjectValue == nil {
		return rnq, fmt.Errorf("no Object in NQuad. Input: [%s]", line)
	}
	if !sane(rnq.Subject) || !sane(rnq.Predicate) || !sane(rnq.ObjectId) {
		// Don't format the full line, as it may contain sensitive information
		return rnq, fmt.Errorf("NQuad failed sanity check. Subject: %q, Predicate: %q, ObjectId: %q",
			rnq.Subject, rnq.Predicate, rnq.ObjectId)
	}

	return rnq, nil
}

// parseFunction parses uid(<var name>) and returns
// uid(<var name>) after striping whitespace if any
func parseFunction(it *lex.ItemIterator) (string, error) {
	item := it.Item()
	s := item.Val

	it.Next()
	if item = it.Item(); item.Typ != itemLeftRound {
		return "", fmt.Errorf("expected '(', found: %s", item.Val)
	}

	it.Next()

View on GitHub (pinned to 759e242be6)