sql.json配置文件参数parallel_num导致查询过滤条件失效

在1.0版本，经过分析日志信息，增加配置文件sql.json, 将默认值"parallel_num": 1调整为"parallel_num": 10, 提高查询性能，
但是会导致查询过滤条件失效，只有将parallel_num的参数值改回1，才能正常过滤数据。
请问如何配置参数，实现并行处理？
/config/online_config/bizs/default/0/sql.json
{
"catalog_name": "",
"db_name": "",
"output_format": "",
"parallel_num": 1,
"sql_timeout": 1000,
"log_search_info_threshold": 0,
"timeout_factor": 0.5,
"sub_graph_timeout_factor": 0.9,
"sub_graph_thread_limit": 10,
"main_graph_thread_limit": 5,
"lack_result_enable": false,
"iquan_plan_prepare_level": "jni.post.optimize",
"iquan_plan_cache_enable": false,
"inner_docid_optimize_enable": false,
"parallel_tables": [],
"summary_tables": [],
"sql_runtime_task_queue": "",
"logic_tables": {
"tables": []
},
"layer_tables": {
"layer_tables": []
},
"remote_tables": {
"tables": []
},
"external_table_config": {
"gig_config": {
"subscribe": {
"cm2": {
"zk_host": "",
"zk_path": "",
"enable_cluster_biz_search": false,
"sub_master_only": false,
"sub_protocol_version": -1,
"cm2_part": [],
"cm2_part_no_topo_info": [],
"cm2_group": [],
"cm2_all_group_topo": false
},
"vip": {
"jmenv_dom": "",
"services": []
},
"istio": {
"istio_host": "",
"biz_clusters": [],
"xdsCallbackQueueSize": 1000,
"enable_host_transform": true,
"cacheFile": "./local_cache/istio_cluster_cache",
"readCache": true,
"xdsClientWorkerThreadNumber": 1,
"writeCache": true,
"writeInterval": 300
},
"local": [],
"cm2_configs": [],
"istio_configs": [],
"allow_empty_sub": true
},
"flow_control": {}
},
"service_config": {},
"table_config": {}
},
"enable_scan_timeout": true,
"async_scan_concurrency": 18446744073709551615,
"target_watermark_timeout_ratio": 0.5,
"cte_opt_ver": 0,
"need_print_slow_log_factory": -1.0,
"need_print_error_log": false,
"enable_turbojet": false,
"db_name_alias": {},
"table_name_alias": {},
"sql_agg_plugin_config": {
"modules": []
},
"sql_tvf_plugin_config": {
"modules": [],
"tvf_profiles": []
},
"iquan_client_config": {
"debug_flag": false,
"cache_config": {
"jni.post.optimize": {
"initial_capacity": 1024,
"concurrency_level": 8,
"maximum_size": 104857600
},
"rel.post.optimize": {
"initial_capacity": 1024,
"concurrency_level": 8,
"maximum_size": 4096
},
"rel.transform": {
"initial_capacity": 1024,
"concurrency_level": 8,
"maximum_size": 4096
},
"sql.parse": {
"initial_capacity": 1024,
"concurrency_level": 8,
"maximum_size": 4096
}
}
},
"iquan_jni_config": {
"table_config": {
"summary_suffix": "_summary_"
}
},
"iquan_warmup_config": {
"thread_number": 3,
"warmup_second_interval": 30,
"warmup_query_number": 1000000,
"warmup_file_path": "",
"warmup_log_name": "sql_warmup"
},
"iquan_kmon_config": {
"service_name": "",
"tenant_name": "default",
"flume_address": "127.0.0.1:4141",
"auto_recycle": true,
"global_tags": ""
},
"authentication_config": {
"enable": false,
"token_pairs": {}
},
"swift_writer_config": {
"swift_client_config": "",
"table_read_write_config": {}
},
"table_writer_config": {
"zone_names": [],
"allow_follow_write": false
}
}

Start with /config/online_config/bizs/default/0/sql.json and trace how parallel_num is consumed during SQL query execution. Reproduce the difference between values 1 and 10 using a query with filtering conditions; done means parallel processing preserves filtering, with the supported configuration or limitation made clear.

Written by the indexing model from the issue text.