fix(dwh): align adapter sampling contract

harness/tht/db/execute.py

@@ -4,39 +4,19 @@ from sqlalchemy import Engine
 
 from tht.execute import ExecResult, PlanSummary, explain as _explain, run_controlled
 
-DEFAULT_LIMIT = 10
 DEFAULT_TIMEOUT_MS = 30_000
 
 
-def run_query(engine: Engine, sql: str, *, limit: int | None = None) -> ExecResult:
+def run_query(engine: Engine, sql: str, *, limit: int) -> ExecResult:
+    if limit <= 0:
+        raise ValueError("limit must be a positive integer")
     return run_controlled(
         engine,
         sql,
-        limit=DEFAULT_LIMIT if limit is None else limit,
+        limit=limit,
         timeout_ms=DEFAULT_TIMEOUT_MS,
     )
 
 
 def explain(engine: Engine, sql: str) -> PlanSummary:
     return _explain(engine, sql, timeout_ms=DEFAULT_TIMEOUT_MS)
-
-
-def sample_column(
-    engine: Engine, schema: str, table: str, column: str, *, limit: int
-) -> list[object]:
-    quote = engine.dialect.identifier_preparer.quote
-    sql = (
-        f"SELECT {quote(column)} FROM {quote(schema)}.{quote(table)} "
-        f"WHERE {quote(column)} IS NOT NULL"
-    )
-    return [row[0] for row in run_query(engine, sql, limit=limit).rows]
-
-
-def distinct_values(engine: Engine, schema: str, table: str, column: str) -> list[object]:
-    quote = engine.dialect.identifier_preparer.quote
-    sql = (
-        f"SELECT DISTINCT {quote(column)} FROM {quote(schema)}.{quote(table)} "
-        f"WHERE {quote(column)} IS NOT NULL ORDER BY {quote(column)}"
-    )
-    result = run_controlled(engine, sql, limit=1000, timeout_ms=DEFAULT_TIMEOUT_MS)
-    return [row[0] for row in result.rows]