HumanSignal/label-studio · error · ValidationError

Label config contains non-unique names:

Error message

Label config contains non-unique names: 

What it means

Label config tag names (name="...") must be unique. After schema validation, validate_label_config collects all name="..." values via regex and raises this ValidationError if the set of names is smaller than the list (duplicates exist). Note the FIXME: names written with spaces like 'name =' are not matched.

Source

Thrown at label_studio/core/label_config.py:127

    try:
        config, cleaned_config_string = parse_config_to_json(config_string)
        jsonschema.validate(config, _LABEL_CONFIG_SCHEMA_DATA)
    except (etree.ParseError, ValueError) as exc:
        raise ValidationError(str(exc))
    except jsonschema.exceptions.ValidationError as exc:
        # jsonschema4 validation error now includes all errors from "anyOf" subschemas
        # check https://python-jsonschema.readthedocs.io/en/latest/errors/#jsonschema.exceptions.ValidationError.context
        # we pick the first failed schema and show only its error message
        error_message = exc.context[0].message if len(exc.context) else exc.message
        error_message = 'Validation failed on {}: {}'.format(
            '/'.join(map(str, exc.path)), error_message.replace('@', '')
        )
        raise ValidationError(error_message)

    # unique names in config # FIXME: 'name =' (with spaces) won't work
    all_names = re.findall(r'(?:^|[^\w])name="([^"]*)"', cleaned_config_string)
    if len(set(all_names)) != len(all_names):
        raise ValidationError('Label config contains non-unique names: ' + ', '.join(all_names))

    # toName points to existent name
    names = set(all_names)
    toNames = re.findall(r'toName="([^"]*)"', cleaned_config_string)
    for toName_ in toNames:
        for toName in toName_.split(','):
            if toName not in names:
                raise ValidationError(f'toName="{toName}" not found in names: {sorted(names)}')

    # Tag attribute validation (e.g. Video playback speed) via SDK
    try:
        li = LabelInterface(config_string)
        if hasattr(li, '_tag_attribute_validation'):
            li._tag_attribute_validation()
    except LabelStudioValidationErrorSentryIgnored as exc:
        raise ValidationError(str(exc))

View on GitHub (pinned to 0b49e9b539)

Solutions

1. Find duplicate names in your config (search name=" occurrences) and rename all but one uniquely.
2. Use descriptive unique names per tag, e.g. name="text1", name="text2".
3. Remember to update corresponding toName="..." references after renaming.
4. Note: 'name =' with spaces before '=' isn't detected by this check — keep attribute syntax as name="value".

Example fix

<!-- before -->
<Text name="text" value="$article"/>
<Text name="text" value="$summary"/>
<!-- after -->
<Text name="article" value="$article"/>
<Text name="summary" value="$summary"/>

import re
defens_names = re.findall(r'(?:^|[^\w])name="([^"]*)"', config)
dups = {n for n in defens_names if defens_names.count(n) > 1}
if dups:
    raise ValueError(f'Duplicate tag names: {dups}')

try:
    validate_label_config(config)
except ValidationError as e:
    if 'non-unique names' in str(e):
        return Response({'error': str(e), 'hint': 'Rename duplicate tag name attributes'}, status=400)
    raise

When it happens

Trigger: A label config containing two or more tags with the same name="..." value, passed to validate_label_config (e.g. saving project labeling config via API or UI).

Common situations: Duplicating a tag block (e.g. two <Text name="text"> or two <Choices name="sentiment">) when composing complex configs; merging configs from templates without renaming.

Related errors

• toName="{toName}" not found in names: {sorted(names)}
• Validation failed on {}: {}
• Your label config has more than one data key and direct file
• {'predictions': prediction_errors}
• "{data_key}" key is expected in task data

AI-assisted analysis of HumanSignal/label-studio@0b49e9b539 (2026-08-29). Data as JSON: /api/errors/f5b7dced6fddcec6. Report an issue: GitHub.