{
  "context": "This criteria evaluates how well the engineer uses ConnectorX's metadata retrieval capabilities to inspect database schemas without loading actual data. The focus is on proper use of ConnectorX's read_sql function for schema inspection.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Uses read_sql function",
      "description": "Uses connectorx.read_sql() function as the primary method for retrieving data from the database rather than implementing custom database connection and query execution logic",
      "max_score": 30
    },
    {
      "name": "Schema extraction",
      "description": "Extracts schema information by examining the DataFrame structure using df.dtypes for data types and df.columns for column names after executing query",
      "max_score": 25
    },
    {
      "name": "Type string conversion",
      "description": "Converts pandas dtype objects to string representations for the return dictionary (e.g., using str(dtype) or dtype.name)",
      "max_score": 20
    },
    {
      "name": "Efficient metadata retrieval",
      "description": "Optimizes queries to minimize data transfer using SQL LIMIT clause or similar technique (e.g., LIMIT 0 or LIMIT 1) when only schema information is needed",
      "max_score": 15
    },
    {
      "name": "Schema comparison implementation",
      "description": "Implements compare_schemas by retrieving both schemas using get_schema and identifying common/unique columns through set operations or dictionary key comparisons",
      "max_score": 10
    }
  ]
}