dgraph-io/dgraph · error

Unable to peek: %v

Error message

Unable to peek: %v

What it means

parseNamespace in the schema lexer/parser peeks 2 tokens ahead to read a namespace like [0x1] at the start of a schema block. If the token iterator cannot peek that far (EOF or iterator exhausted), the parser wraps the underlying peek error with 'Unable to peek: %v'. It means the schema text ended before a namespace declaration could be read.

Source

Thrown at schema/parse.go:584

		if it.Item().Typ == itemExclamationMark {
			it.Next()
		}
	}

	if it.Item().Typ != itemNewLine {
		return nil, it.Item().Errorf("Expected new line after field declaration. Got %v",
			it.Item().Val)
	}

	glog.Warningf("Type declaration for type %s includes deprecated information about field type "+
		"for field %s which will be ignored.", typeName, x.ParseAttr(field.Predicate))
	return field, nil
}

func parseNamespace(it *lex.ItemIterator) (uint64, error) {
	nextItems, err := it.Peek(2)
	if err != nil {
		return 0, errors.Errorf("Unable to peek: %v", err)
	}
	if nextItems[0].Typ != itemNumber || nextItems[1].Typ != itemRightSquare {
		return 0, errors.Errorf("Typed oes not match the expected")
	}
	ns, err := strconv.ParseUint(nextItems[0].Val, 0, 64)
	if err != nil {
		return 0, err
	}
	it.Next()
	it.Next()
	// We have parsed the namespace. Now move to the next item.
	if !it.Next() {
		return 0, errors.Errorf("No schema found after namespace. Got: %v", nextItems[0])
	}
	return ns, nil
}

// ParsedSchema represents the parsed schema and type updates.

View on GitHub (pinned to 759e242be6)

Solutions

1. Provide the complete schema text, including '<predicate>: <type> .' entries after the [namespace] section
2. Check the schema source file for truncation (verify it ends with a newline/type declarations)
3. If namespaces are not needed, omit the '[' bracket entirely instead of emitting a dangling one
4. Log the raw schema string before Parse to confirm what is actually being parsed

Example fix

// before
schema := "name: string @index(exact) .\n[0x1]" // truncated after namespace
// after
schema := "name: string @index(exact) .\n[0x1]\nname: string @index(exact) ."

if strings.TrimSpace(schemaStr) == "" || strings.HasSuffix(strings.TrimSpace(schemaStr), "[") {
    return fmt.Errorf("schema text truncated before namespace")
}
err := schema.Parse(schemaStr, 0, 1)

if _, err := schema.Parse(s, 0, 1); err != nil {
    if strings.Contains(err.Error(), "Unable to peek") {
        log.Fatalf("truncated schema: %v\nraw=%q", err, s)
    }
    return err
}

When it happens

Trigger: Calling schema.Parse / ParseWithNamespace on schema DQL whose text terminates right after '[' or truncates before '<predicate>: type' following the [ns] token, e.g. an empty or cut-off schema string.

Common situations: Schema file truncated during upload/copy; passing an empty string; manually editing schema and deleting the body after the namespace bracket; tooling generating schema programmatically that stops emitting after the opening bracket.

Related errors

• Typed oes not match the expected
• No schema found after namespace. Got: %v
• While parsing namespace:
• NQuad failed sanity check. Subject: %q, Predicate: %q, Objec
• expected '(', found: %s

AI-assisted analysis of dgraph-io/dgraph@759e242be6 (2026-09-01). Data as JSON: /api/errors/d27034e016e6abd8. Report an issue: GitHub.