Question

我有一个SQL查询，我想优化以改善结果。

查询：

SELECT 
    twm.id AS id, 
    twm.fecha_id AS fecha,
    twm.nombre_busqueda AS busquedas, 
    twm.tipo_mencion_leyenda AS tipo_mencion, 
    twm.sentimiento AS sentimiento,
    CASE 
        WHEN sum(twm.menciones) IS null THEN 0 
        ELSE sum(twm.menciones) 
    END AS menciones,
    twm.tipo_source_nombre AS sources, 
    twm.tipo_de_site_leyenda AS typeof
FROM 
    dashboards_tablaagregadawebdiaria twm
WHERE 
    twm.corporativa = FALSE 
    AND twm.cuenta_id = 115 
    AND twm.tipo_buzz_id = 1    
    AND twm.fecha_id >= 20150831 
    AND twm.fecha_id <= 20150929
GROUP BY 
    twm.id,fecha, 
    busquedas, 
    tipo_mencion, 
    sentimiento, 
    sources, 
    typeof

解释分析的输出如下：

GroupAggregate  (cost=9052.86..9528.69 rows=13595 width=50) (actual   time=45.070..56.073 rows=15006 loops=1)
  ->  Sort  (cost=9052.86..9086.85 rows=13595 width=50) (actual time=45.061..47.941 rows=15006 loops=1)
    Sort Key: id, fecha_id, nombre_busqueda, tipo_mencion_leyenda, sentimiento, tipo_source_nombre, tipo_de_site_leyenda
    Sort Method: external merge  Disk: 1008kB
    ->  Bitmap Heap Scan on dashboards_tablaagregadawebdiaria twm  (cost=910.51..7654.01 rows=13595 width=50) (actual time=3.491..31.813 rows=15006 loops=1)
          Recheck Cond: (cuenta_id = 115)
          Filter: ((NOT corporativa) AND (fecha_id >= 20150831) AND (fecha_id <= 20150929) AND (tipo_buzz_id = 1))
         ->  Bitmap Index Scan on dashboards_tablaagregadawebdiaria_cuenta_id  (cost=0.00..907.12 rows=21175 width=0) (actual time=3.127..3.127 rows=21474 loops=1)
                Index Cond: (cuenta_id = 115)
Total runtime: 57.057 ms

模型如下：

CREATE TABLE dashboards_tablaagregadawebdiaria
(
  id serial NOT NULL,
  busqueda_id integer NOT NULL,
  fecha_id integer NOT NULL,
  calidad_id integer NOT NULL,
  tipo_de_site_id integer NOT NULL,
  tipo_buzz_id integer NOT NULL,
  visibilidad_id integer NOT NULL,
  menciones integer,
  tipo_menciones_id integer,
  sentimiento double precision NOT NULL,
  termino_encontrado character varying(200),
  pais_id integer,
  idioma_id integer,
  source_id integer NOT NULL,
  cuenta_id integer,
  corporativa boolean,
  nombre_busqueda character varying(150),
  tipo_source_id integer,
  tipo_source_nombre character varying(100),
  calidad_leyenda character varying(200),
  visibilidad_leyenda character varying(200),
  tipo_buzz_leyenda character varying(200),
  tipo_mencion_leyenda character varying(200),
  tipo_de_site_leyenda character varying(200),
  CONSTRAINT dashboards_tablaagregadawebdiaria_pkey PRIMARY KEY (id),
  CONSTRAINT busqueda_id_refs_id_a3e56469 FOREIGN KEY (busqueda_id)
  REFERENCES busquedas_busqueda (id) MATCH SIMPLE
  ON UPDATE NO ACTION ON DELETE NO ACTION DEFERRABLE INITIALLY DEFERRED,
  CONSTRAINT calidad_id_refs_id_60e9ae3b FOREIGN KEY (calidad_id)
  REFERENCES busquedas_calidad (id) MATCH SIMPLE
  ON UPDATE NO ACTION ON DELETE NO ACTION DEFERRABLE INITIALLY DEFERRED,
  CONSTRAINT cuenta_id_refs_id_1473e5b2 FOREIGN KEY (cuenta_id)
  REFERENCES usuarios_cuenta (id) MATCH SIMPLE
  ON UPDATE NO ACTION ON DELETE NO ACTION DEFERRABLE INITIALLY DEFERRED,
  CONSTRAINT fecha_id_refs_id_614e1caa FOREIGN KEY (fecha_id)
  REFERENCES dashboards_tiempo (id) MATCH SIMPLE
  ON UPDATE NO ACTION ON DELETE NO ACTION DEFERRABLE INITIALLY DEFERRED,
  CONSTRAINT idioma_id_refs_id_dea0e30f FOREIGN KEY (idioma_id)
  REFERENCES usuarios_idioma (id) MATCH SIMPLE
  ON UPDATE NO ACTION ON DELETE NO ACTION DEFERRABLE INITIALLY DEFERRED,
  CONSTRAINT pais_id_refs_id_c3a64f5e FOREIGN KEY (pais_id)
  REFERENCES usuarios_pais (id) MATCH SIMPLE
  ON UPDATE NO ACTION ON DELETE NO ACTION DEFERRABLE INITIALLY DEFERRED,
  CONSTRAINT source_id_refs_id_b20099a3 FOREIGN KEY (source_id)
  REFERENCES busquedas_source (id) MATCH SIMPLE
  ON UPDATE NO ACTION ON DELETE NO ACTION DEFERRABLE INITIALLY DEFERRED,
  CONSTRAINT tipo_buzz_id_refs_id_06719001 FOREIGN KEY (tipo_buzz_id)
  REFERENCES dashboards_tipo_buzz (id) MATCH SIMPLE
  ON UPDATE NO ACTION ON DELETE NO ACTION DEFERRABLE INITIALLY DEFERRED,
  CONSTRAINT tipo_de_site_id_refs_id_74d8ca3e FOREIGN KEY (tipo_de_site_id)
  REFERENCES dashboards_tipo_site (id) MATCH SIMPLE
  ON UPDATE NO ACTION ON DELETE NO ACTION DEFERRABLE INITIALLY DEFERRED,
  CONSTRAINT tipo_menciones_id_refs_id_6aa6306a FOREIGN KEY    (tipo_menciones_id)
  REFERENCES dashboards_tipo_mencion (id) MATCH SIMPLE
  ON UPDATE NO ACTION ON DELETE NO ACTION DEFERRABLE INITIALLY DEFERRED,
  CONSTRAINT tipo_source_id_refs_id_38d1acb3 FOREIGN KEY (tipo_source_id)
  REFERENCES busquedas_tipo_source (id) MATCH SIMPLE
  ON UPDATE NO ACTION ON DELETE NO ACTION DEFERRABLE INITIALLY DEFERRED,
  CONSTRAINT visibilidad_id_refs_id_9f3d8b04 FOREIGN KEY (visibilidad_id)
  REFERENCES dashboards_visibilidad_web (id) MATCH SIMPLE
  ON UPDATE NO ACTION ON DELETE NO ACTION DEFERRABLE INITIALLY DEFERRED,
  CONSTRAINT dashboards_tablaagregadawebdiaria_busqueda_id_5df9ff43_uniq   UNIQUE (busqueda_id, fecha_id, calidad_id, tipo_de_site_id, tipo_buzz_id,   visibilidad_id, tipo_menciones_id, termino_encontrado, pais_id, idioma_id,   sentimiento, source_id)
)
WITH (
  OIDS=FALSE
);
ALTER TABLE dashboards_tablaagregadawebdiaria
  OWNER TO postgres;

提高查询时间的任何帮助或建议？我已经创建了跟随索引，但还不够。

CREATE INDEX dashboards_tablaagregadawebdiaria_cuenta_id
 ON dashboards_tablaagregadawebdiaria
  USING btree
  (cuenta_id);

由于

Answer 1

查询中只有一个表格带有std::cout子句。最佳指数是复合指数：

std::string

优化PostgreSQL中的SQL查询（选择正确的索引）

1 个答案: