How to count and order by number of incoming edges
I am trying to get get the ten people who have starred in the most movies and get the name and the number of movies in decending order. I am using the IMDB dataset.
This is what I have so far:
arangosh [_system]> var Graph = require("org/arangodb/graph").Graph;
arangosh [_system]> var db = require("org/arangodb").db;
arangosh [_system]> statement = db._createStatement({query: 'FOR vert IN imdb_vertices FILTER vert.type == "Person" LET edge_count = (LENGTH(EDGES(imdb_edges, vert, "outbound", [{type: "ACTS_IN"}]))) RETURN {"name": vert.name, "count": edge_count}'})
[object ArangoStatement]
arangosh [_system]> cursor = statement.execute()
[object ArangoQueryCursor]
arangosh [_system]> cursor.next()
{
"name" : "Stephanie Faracy",
"count" : 0
}
How can I do that with AQL?
Without an index: The following will return the Top-10:
arangosh [_system]> db._createStatement({query: 'FOR vert IN imdb_vertices FILTER vert.type == "Person" LET edge_count = (LENGTH(EDGES(imdb_edges, vert, "outbound", [{"type": "Role", "$label": "ACTS_IN"}]))) SORT edge_count DESC LIMIT 10 RETURN {"name": vert.name, "count": edge_count}'}).execute().toArray()
[
{
"name" : "Clint Eastwood",
"count" : 148
},
{
"name" : "Claude Jade",
"count" : 142
},
{
"name" : "Samuel L. Jackson",
"count" : 122
},
{
"name" : "Armin Mueller-Stahl",
"count" : 112
},
{
"name" : "Gérard Depardieu",
"count" : 104
},
{
"name" : "Marisa Mell",
"count" : 104
},
{
"name" : "Robert De Niro",
"count" : 104
},
{
"name" : "Bruce Willis",
"count" : 96
},
{
"name" : "Jackie Chan",
"count" : 94
},
{
"name" : "Michael Caine",
"count" : 90
}
]
Basically you can use the "sort" also for variables created with LET. Limit allows you to restrict to the TOP 10. Note that the type in vertex is 'Role' and label is 'ACTS_IN'.
It would be more efficient to add the number to the documents and use a sorted index. But this would require to updating the documents.
arangosh [_system]> c = db._createStatement({query: 'FOR vert IN imdb_vertices FILTER vert.type == "Person" LET edge_count = (LENGTH(EDGES(imdb_edges, vert, "outbound", [{"type": "Role", "$label": "ACTS_IN"}]))) RETURN {"_key": vert._key, "count": edge_count}'}).execute()
[object ArangoQueryCursor]
arangosh [_system]> while (c.hasNext()) { var d = c.next(); db.imdb_vertices.update(d._key, {COUNT: d.count}); }
arangosh [_system]> db.imdb_vertices.ensureSkiplist("COUNT");
arangosh [_system]> x = db._createStatement({query: 'FOR vert in imdb_vertices FILTER vert.COUNT >= 0 SORT vert.COUNT DESC LIMIT 10 RETURN vert'}).execute()
[object ArangoQueryCursor]
链接地址: http://www.djcxy.com/p/78732.html
上一篇: 从列表末尾删除连续的事件
下一篇: 如何通过输入边缘的数量进行计数和排序