Fix tuple parameter handling for SQLAlchemy 2.0 pooled queries

Fixes 'List of argument must consist only of dictionaries' error when
running parallel queries with ODBC-style ? placeholders.

The issue occurs because SQLAlchemy 2.0's text() expects named
parameters (dict) but the code was passing positional parameters (tuples).

Solution:
- Detect tuple parameters in _execute_query_pooled
- Convert tuples to dict with keys p1, p2, p3...
- Replace ? placeholders with :p1, :p2... in the query string
- Pass converted dict to conn.execute()

This maintains backward compatibility with single-connection mode
which uses pyodbc and handles tuple parameters natively.

Affected queries:
- table_exists (2 params: schema, table)
- get_columns (2 params: schema, table)
- get_primary_keys (2 params: schema, table)

src/drt/database/executor.py

@@ -75,6 +75,24 @@ class QueryExecutor:
             raise ValueError("Only SELECT queries are allowed (READ ONLY)")
 
         try:
+            # Convert tuple params to dict for SQLAlchemy 2.0 compatibility
+            if isinstance(params, tuple):
+                # Map positional params to p1, p2, p3...
+                params_dict = {f"p{i}": val for i, val in enumerate(params, 1)}
+                # Replace ? placeholders with :p1, :p2... by iterating
+                query_parts = []
+                last_pos = 0
+                param_index = 1
+                for char_pos, char in enumerate(query):
+                    if char == '?':
+                        query_parts.append(query[last_pos:char_pos])
+                        query_parts.append(f':p{param_index}')
+                        last_pos = char_pos + 1
+                        param_index += 1
+                query_parts.append(query[last_pos:])
+                query = ''.join(query_parts)
+                params = params_dict
+            
             with self._engine.connect() as conn:
                 result = conn.execute(text(query), params or {})
                 df = pd.DataFrame(result.all(), columns=result.keys())