pgrouting学习记录

1 数据库尝试

路径规划的算法 pgr_dijkstra，pgr_astar, pgr_ksp,pgr_withpoints_参考1,pgr_withpoints_参考2

其中pgr_ksp
是dijkstra算法和yen算法结合，获取两点之间的k条最短路径

• pgr_dijkstra 算法

SELECT gid, geom, node, cost, source, target, 
    ST_Reverse(geom) AS flip_geom FROM 
	pgr_dijkstra(
	'SELECT gid AS id,
		source, target,
		cost_len as cost, rcost_len as reverse_cost
	FROM r1',
	3, 7,
	directed := FALSE
),r1 WHERE edge = gid ORDER BY seq;

• pgr_ksp

参数解释：3，7，3 分别为source, target, 返回的几条数据

select * from	
 pgr_ksp('SELECT gid as id, source, target, cost_len AS cost , cost_len AS revers_cost
 FROM r1',
						3, 7,3,false ),r1
     WHERE edge = gid ORDER BY seq

• pgr_withpoints (edges_sql, points_sql, start_vid, end_vid)

pgr_withPoints和 pgr_dijkstra一样也是最短路径规划方法，输出结果也基本一致，但是两者的输入不一样。pgr_dijkstra需要输入的是路网节点编号，也就是说只能从道路的端点算起，而pgr_withPoints可以从道路的中间位置选择起点和终点。

• 参数说明：
  edges_sql是从路网表选择出需要参与计算的路径的Sql字符串;
  points_sql是从POI表（下文会说明）选择出途径点的Sql字符串;
  start_vid是起点的节点id，正值为路网节点，负值为POI表数据;
  end_vid为终点节点，同样正值为路网节点，负值为POI表数据。
  注意在edges_sql中需要指定cost和reverse_cost（去的花费和来的花费），函数通过这两个值的正负情况来获知道路的通过性（单行、双向、不通）

• pgr_astar

select * from
 pgr_astar('SELECT gid as id, source::int, target::int, cost_len::float AS cost,
		   x1,y1,x2,y2 FROM r1',
			array[2,3], array[7,26] ,false),r1
     WHERE edge = gid ORDER BY seq

• pgr_bdastar

SELECT gid, geom, node, cost, source, target, 
    ST_Reverse(geom) AS flip_geom FROM 
 pgr_bdAstar('SELECT gid as id, source, target,
  cost_len AS cost,x1,y1,x2,y2
 FROM r1',
						2, 7 ,false),r1
     WHERE edge = gid ORDER BY seq