具有 3 个分离度级别的表 [英] Table with 3 levels of degree of separation

问题描述

我正在寻找一个查询来在 MySQL 中创建一个具有两个用户之间分离程度的表.我已经找到了分离度查询.但是，如果我理解正确，这将导致与共同朋友推荐的好友列表.我要找的东西略有不同.

我有一个包含用户之间朋友"的表格(不包含 1 到 2 和 2 到 1 之类的重复关系).

<块引用>

friends (id, initiator_user_id, friend_user_id, is_confirmed)

我正在尝试创建一个表格，其中包含朋友、朋友的朋友和 FoFoF 之间的所有关系.像这样:

<块引用>

relation_degrees (id, first_user_id, second_user_id, relation_degree)

因此relation_degree 列仅包含值1(朋友)、2(FoF)和3(FoFoF).

我能够在 Excel 中完成它，但是我的朋友们存储在一个矩阵中，这使得 IMO 的计算更容易一些.我希望有人能够给我一个提示，让我在 MySQL 中做同样的事情.

谢谢！！

<小时>

在 Fluffeh 的帮助下，我找到了以下解决我的问题的方法.

1. 我将两个方向的关系(如 1-2 & 2-1，所以没有确认列)存储在名为 degree_one 的表中
2. 然后我使用了来自 fluffeh 的第一级和第二级查询来制作带有 first & 的表.二级关系.我添加了一个 WHERE 用户 <> Friend 语句来过滤关系(我想这是来自 fluffeh 的三级关系查询无法正常工作的原因之一)

<块引用>

 `创建表 degree_two选择mb.user 作为用户，mb.friend 作为朋友，min(mb.rel) 作为relation_degree

来自(选择1 作为相对，fr1.用户，fr1.朋友从degree_one fr1

 联合所有选择2 作为相对，fr2.用户，fr3.朋友从degree_one fr2左外连接 degree_one fr3在 fr2.Friend=fr3.User) mb哪里用户 <>朋友通过...分组mb.用户，mb.朋友

3. 然后我用这个表做了几乎相同的查询.select 语句在 degree_two 表上，但外连接仍然来自 degree_one 表.

<块引用>

 创建表 degree_three选择mb.user 作为用户，mb.friend 作为朋友，min(mb.relation_degree) 作为relation_degree从(选择fr1.relation_degree，fr1.用户，fr1.朋友从degree_two fr1联合所有选择3 作为相对，fr2.用户，fr3.朋友从degree_two fr2左外连接 degree_one fr3在 fr2.Friend=fr3.User) mb其中`用户` <>`朋友`通过...分组mb.用户，mb.朋友

这是一种解决方法，但它为我提供了所需的输出.我仍然想知道为什么来自 fluffeh 的查询不能正常工作，因为我真的想要一个单一的查询作为解决方案.我将继续玩弄查询...我希望有人可以帮助我将这些查询合并为一个.

解决方案

这个查询给了我想要的解决方案结果:

创建表 degree_two选择mb.user 作为用户，mb.friend 作为朋友，min(mb.rel) 作为relation_degree从(选择1 作为相对，fr1.用户，fr1.朋友从degree_one fr1联合所有选择2 作为相对，fr2.用户，fr3.朋友从degree_one fr2左外连接 degree_one fr3在 fr2.Friend=fr3.User其中 fr2.user <>fr3.朋友联合所有选择3 作为相对，fr4.user,fr6.朋友从(degree_one fr4左外连接 degree_one fr5在 fr4.friend=fr5.user和 fr4.user <>fr5.朋友左外连接 degree_one fr6在 fr5.friend = fr6.user和 fr5.user <>fr6.朋友)其中 fr6.friend 不是 NULL) mb通过...分组mb.用户，mb.朋友

感谢 Fluffeh 的帮助！

I'm looking for a query to create a table in MySQL with the degree of seperation between two users. I already found Degrees of Separation Query.But that will, if i understand correctly, result in a recommended friendlist with mutual friends. What i'm looking for is slightly different.

I have a table with "friends" between users (contains no duplicate relations like 1 to 2 & 2 to 1).

friends (id, initiator_user_id, friend_user_id, is_confirmed)

What i am trying to create is a table with all relations between friends, friends of friends and FoFoF. like this:

relation_degrees (id, first_user_id, second_user_id, relation_degree)

so the relation_degree column only contains the value 1 (friends), 2 (FoF) and 3 (FoFoF).

I was able to do it in Excel, but there my friends where stored in a matrix, which make calculations IMO a little bit easier. I hope somebody will be able to give me a hint to do the same in MySQL.

Thanks!!

---

edit: with the help from Fluffeh i found the following solution to my problem.

1. i stored the relations in both directions( like 1-2 & 2-1, so without the confirmation column) in table called degree_one
2. Then i used the query for degree one and two from fluffeh to make table with first & second degree relations. I added a WHERE user <> Friend statement to filter the relations (i guess this is one of the reasons why the query from fluffeh for the third degree relations isn't working correctly)

   `Create table degree_two
   select
   mb.user as User,
   mb.friend as Friend,
   min(mb.rel)  as relation_degree

from ( select 1 as rel, fr1.User, fr1.Friend from degree_one fr1

          union all

      select
          2 as rel,
           fr2.User,
           fr3.Friend
       from
           degree_one fr2
               left outer join degree_one fr3
                   on fr2.Friend=fr3.User


                ) mb

   Where user <> friend



 group by
      mb.User,
      mb.Friend

3. Then i used this table to do almost the same query. the select statement is on the degree_two table but the outer join is still from the degree_one table.

   Create table degree_three
    select
        mb.user as User,
        mb.friend as Friend,
        min(mb.relation_degree)  as relation_degree
    from
        (
             select
                fr1.relation_degree,
                fr1.User,
                fr1.Friend
            from
                degree_two fr1


          union all

      select
            3 as rel,
            fr2.User,
            fr3.Friend
        from
            degree_two fr2
                left outer join degree_one fr3
                    on fr2.Friend=fr3.User


                ) mb

Where `user` <> `friend`

group by
    mb.User,
    mb.Friend

It is kind of a work around but it gives me the desired output. I'm still wondering why the query from fluffeh doesn't work correctly, because i really want a single query as solution. i will continue fooling around with the query... I hope that somebody can help me merging these query into one.

解决方案

This query gave me the desired solution result:

Create table degree_two
select
    mb.user as User,
    mb.friend as Friend,
    min(mb.rel)  as relation_degree
from
    (
           select
            1 as rel,
            fr1.User,
            fr1.Friend
        from
            degree_one fr1

            union all

        select
            2 as rel,
            fr2.User,
            fr3.Friend
        from
            degree_one fr2
                left outer join degree_one fr3
                    on fr2.Friend=fr3.User
                    where fr2.user <> fr3.friend

         union all       

         select 
            3 as rel, 
            fr4.user, 
            fr6.Friend
        from 
           (degree_one fr4 
                left outer join  degree_one fr5 
                    on fr4.friend=fr5.user
                    and fr4.user <> fr5.friend

                    left outer join  degree_one fr6
                    on fr5.friend = fr6.user
                   and fr5.user <> fr6.friend)
           where fr6.friend  IS NOT NULL           


                  ) mb

group by
    mb.User,
    mb.Friend

Thanks Fluffeh for your help!

这篇关于具有 3 个分离度级别的表的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！