Combine "Query full " and "Query results" to reduce the cost of queries.

[Smallhi]

2024-06-02 14:43:49,927:INFO:superset.common.query_object:query cache dict {'columns': ['ds', 'os', 'brand'], 'extras': {'having': '', 'where': "(ds ='20240601')", 'time_grain_sqla': 'P1D'}, 'filter': [], 'granularity': None, 'inner_from_dttm': None, 'inner_to_dttm': None, 'is_rowcount': False, 'is_timeseries': False, 'metrics': [], 'order_desc': True, 'orderby': [], 'row_limit': 1000, 'row_offset': 0, 'series_columns': [], 'series_limit': 0, 'series_limit_metric': None, 'time_shift': None, 'datasource': '111__table', 'extra_cache_keys': [], 'rls': [], 'changed_on': datetime.datetime(2024, 5, 28, 13, 33, 22), 'result_type': <ChartDataResultType.RESULTS: 'results'>},cache key = c363446f70e39f9854bb24baf0deffe7
2024-06-02 14:43:49,928:INFO:superset.common.query_object:query cache dict {'columns': ['ds', 'os', 'brand'], 'extras': {'having': '', 'where': "(ds ='20240601')", 'time_grain_sqla': 'P1D'}, 'filter': [], 'granularity': None, 'inner_from_dttm': None, 'inner_to_dttm': None, 'is_rowcount': False, 'is_timeseries': False, 'metrics': [], 'order_desc': True, 'orderby': [], 'row_limit': 1000, 'row_offset': 0, 'series_columns': [], 'series_limit': 0, 'series_limit_metric': None, 'time_shift': None, 'datasource': '111__table', 'extra_cache_keys': [], 'rls': [], 'changed_on': datetime.datetime(2024, 5, 28, 13, 33, 22), 'result_type': <ChartDataResultType.FULL: 'full'>},cache key = 7fb33d0e21aab1a4b1a79d8c845742e2

when user click "update chart", there are two request, query full and query result, however, for backend db, the query sql is same. which can increase query cost.