prompt_create_validator

Prompt to create a Pointblank Validator for a loaded DataFrame.

Arguments

df_id: string

ID of the DataFrame to validate.

Provide a value matching the following JSON schema: {“type”:“string”}. Encode non-string values as JSON.

validator_id: string

Optional ID for the Validator. If not provided, a new ID will be generated.

Provide a value matching the following JSON schema: {“anyOf”:[{“type”:“string”},{“type”:“null”}]}. Encode non-string values as JSON.

table_name: string

Optional name for the table within Pointblank reports.

Provide a value matching the following JSON schema: {“anyOf”:[{“type”:“string”},{“type”:“null”}]}. Encode non-string values as JSON.

validator_label: string

Optional descriptive label for the Validator.

Provide a value matching the following JSON schema: {“anyOf”:[{“type”:“string”},{“type”:“null”}]}. Encode non-string values as JSON.

thresholds_dict_example: string

Example thresholds for validation failures.

Provide a value matching the following JSON schema: {“anyOf”:[{“additionalProperties”:{“anyOf”:[{“type”:“integer”},{“type”:“number”}]},“type”:“object”},{“type”:“null”}]}. Encode non-string values as JSON.

Prompt Text

NoteAssistant

Once your data is loaded (using its `df_id`), I can create a 'Validator' object to define data quality checks.

NoteUser message

Please call `create_validator` using the `df_id` of your loaded data (e.g., 'df_default').
You can optionally provide:
- `validator_id` (e.g., 'validator_default') to name this validator instance.
- `table_name` (e.g., 'data_table') as a reference name for the data table in reports.
- `validator_label` (e.g., 'Validator') for a descriptive label.
- `thresholds_dict` (e.g., {'warning': 0.05, 'error': 10}) to set global failure thresholds for validation steps.
- Other optional parameters like `actions_dict`, `final_actions_dict`, `brief`, `lang`, `locale` can also be specified if needed.
Make a note of the returned `validator_id` to use when adding validation steps.