Get Data Quality Tool
Get Data Quality Tool
Section titled “Get Data Quality Tool”This tool checks the data quality of one or more catalog tables, or of the tables referenced by a SQL query. Use it to answer “is this data reliable?” before trusting a result. It looks at trust flags, data quality scores, and deprecation warnings, and returns an overall verdict with the specific issues found.
When to use it
Section titled “When to use it”Add this tool to an agent that should validate data before presenting an answer, or that responds to explicit requests like “check data quality” or “is this table trustworthy?”
You provide either a list of table IDs or a SQL query. When you provide a SQL query, you also need to tell the tool which data source it runs against.
Parameters
Section titled “Parameters”| Parameter | Required | Type | Suggested scope | What it does |
|---|---|---|---|---|
table_ids | One of these two | list of integers (max 30) | Agent determined | The tables to check, e.g. [123, 456]. |
sql_query | One of these two | string | Agent determined | A query whose referenced tables should be checked. |
ds_id | With sql_query | integer | User input | The Alation data source ID the query runs against (preferred). |
db_uri | With sql_query | string | Fixed or User input | A database URI, as an alternative to ds_id. |
dq_score_threshold | No | integer 0–100 (default 70) | Fixed | Tables scoring below this are flagged. Higher is stricter. |
output_format | No | "json" (default) or "yaml_markdown" | Fixed | yaml_markdown is more compact when checking many tables. |
default_schema_name | No | string | Fixed | Default schema for unqualified table names in the query, e.g. "public". |
bypassed_dq_sources | No | list of strings | Fixed | Quality sources to skip, e.g. ["trust_flags"]. |
See Configuring tool parameters for what the scope column means.
Example usage
Section titled “Example usage”Check specific tables with a stricter threshold:
{ "table_ids": [123, 456], "dq_score_threshold": 80}Check the tables referenced by a query:
{ "sql_query": "SELECT * FROM orders WHERE order_date > '2024-01-01'", "ds_id": 5}Output
Section titled “Output”An overall status — HIGH DATA QUALITY, LOW DATA QUALITY, or UNKNOWN DATA QUALITY — plus the list of tables that failed checks and any tables where quality could not be determined. If any one table has issues, the overall status is LOW.