MySql PHP从逗号分隔数据(标签)中选择不同值的计数

时间:2014-10-06 11:32:44

标签: php mysql sql

如何从MySql中以逗号分隔值存储的数据中选择不同值的计数?我将使用PHP最终从MySql输出数据。

那里有什么,每个帖子都有标签。所以最后,我正在尝试输出数据,就像stackoverflow对它的标签一样,如下所示:

tag-name x 5

这就是表格中的数据的样子(抱歉内容,但它是食谱的网站)。

"postId"    "tags"                                  "category-code"
"1"         "pho,pork"                              "1"
"2"         "fried-rice,chicken"                    "1"
"3"         "fried-rice,pork"                       "1"
"4"         "chicken-calzone,chicken"               "1"
"5"         "fettuccine,chicken"                    "1"
"6"         "spaghetti,chicken"                     "1"
"7"         "spaghetti,chorizo"                     "1"
"8"         "spaghetti,meat-balls"                  "1"
"9"         "miso-soup"                             "1"
"10"        "chanko-nabe"                           "1"
"11"        "chicken-manchurian,chicken,manchurain" "1"
"12"        "pork-manchurian,pork,manchurain"       "1"
"13"        "sweet-and-sour-pork,pork"              "1"
"14"        "peking-duck,duck"                      "1"

输出

chicken             5 // occurs 5 time in the data above
pork                4 // occurs 4 time in the data above
spaghetti           3 // an so on
fried-rice          2
manchurian          2
pho                 1
chicken-calzone     1
fettuccine          1
chorizo             1
meat-balls          1
miso-soup           1
chanko-nabe         1
chicken-manchurian  1
pork-manchurian     1
sweet-n-sour-pork   1
peking-duck         1
duck                1

我正在尝试select count of all distinct values in there,但由于它是逗号分隔的数据,似乎无法做到这一点。 select distinct无效。

你能想到一个很好的方法,无论是使用mysql还是使用php来获取输出,就像我一样?

5 个答案:

答案 0 :(得分:13)

解决方案

我真的不知道如何将逗号分隔值的水平列表转换为行列表,而不创建包含数字的表,因为您可能使用逗号分隔值。如果你可以创建这个表,这是我的答案:

SELECT 
  SUBSTRING_INDEX(SUBSTRING_INDEX(all_tags, ',', num), ',', -1) AS one_tag,
  COUNT(*) AS cnt
FROM (
  SELECT
    GROUP_CONCAT(tags separator ',') AS all_tags,
    LENGTH(GROUP_CONCAT(tags SEPARATOR ',')) - LENGTH(REPLACE(GROUP_CONCAT(tags SEPARATOR ','), ',', '')) + 1 AS count_tags
  FROM test
) t
JOIN numbers n
ON n.num <= t.count_tags
GROUP BY one_tag
ORDER BY cnt DESC;

返回:

+---------------------+-----+
| one_tag             | cnt |
+---------------------+-----+
| chicken             |   5 |
| pork                |   4 |
| spaghetti           |   3 |
| fried-rice          |   2 |
| manchurain          |   2 |
| pho                 |   1 |
| chicken-calzone     |   1 |
| fettuccine          |   1 |
| chorizo             |   1 |
| meat-balls          |   1 |
| miso-soup           |   1 |
| chanko-nabe         |   1 |
| chicken-manchurian  |   1 |
| pork-manchurian     |   1 |
| sweet-and-sour-pork |   1 |
| peking-duck         |   1 |
| duck                |   1 |
+---------------------+-----+
17 rows in set (0.01 sec)

请参阅sqlfiddle


方案

  1. 我们使用逗号连接所有标签,只创建一个标签列表而不是每行一个
  2. 我们会计算列表中有多少个标签
  3. 我们找到了如何在此列表中获取一个值
  4. 我们发现如何将所有值作为不同的行
  5. 我们统计按其值
  6. 分组的标签

    上下文

    让我们构建您的架构:

    CREATE TABLE test (
        id INT PRIMARY KEY,
        tags VARCHAR(255)
    );
    
    INSERT INTO test VALUES
        ("1",         "pho,pork"),
        ("2",         "fried-rice,chicken"),
        ("3",         "fried-rice,pork"),
        ("4",         "chicken-calzone,chicken"),
        ("5",         "fettuccine,chicken"),
        ("6",         "spaghetti,chicken"),
        ("7",         "spaghetti,chorizo"),
        ("8",         "spaghetti,meat-balls"),
        ("9",         "miso-soup"),
        ("10",        "chanko-nabe"),
        ("11",        "chicken-manchurian,chicken,manchurain"),
        ("12",        "pork-manchurian,pork,manchurain"),
        ("13",        "sweet-and-sour-pork,pork"),
        ("14",        "peking-duck,duck");
    

    连接所有标签列表

    我们将在一行中使用所有代码,因此我们使用GROUP_CONCAT来完成这项工作:

    SELECT GROUP_CONCAT(tags SEPARATOR ',') FROM test;
    

    返回以逗号分隔的所有标记:

      

    PHO,猪肉,炒-米,鸡,油炸米,猪肉,鸡肉,Calzone的,鸡,意大利细面条,鸡,意大利面条,鸡,意大利面条,香肠,意大利面,肉球,味噌汤,相扑火锅,鸡满,鸡,manchurain,猪肉满,猪肉,manchurain,甜和酸味猪肉,猪肉,京鸭,野鸭

    统计所有标签

    要统计所有代码,我们会得到完整代码列表的长度,并且在将,替换为空格后删除完整代码列表的长度。我们加1,因为分隔符在两个值之间。

    SELECT LENGTH(GROUP_CONCAT(tags SEPARATOR ',')) - LENGTH(REPLACE(GROUP_CONCAT(tags SEPARATOR ','), ',', '')) + 1 AS count_tags
    FROM test;
    

    返回:

    +------------+
    | count_tags |
    +------------+
    |         28 |
    +------------+
    1 row in set (0.00 sec)
    

    获取标签列表中的第N个标签

    我们使用SUBSTRING_INDEX函数来获取

    -- returns the string until the 2nd delimiter\'s occurrence from left to right: a,b
    SELECT SUBSTRING_INDEX('a,b,c', ',', 2);
    
    -- return the string until the 1st delimiter, from right to left: c
    SELECT SUBSTRING_INDEX('a,b,c', ',', -1);
    
    -- we need both to get: b (with 2 being the tag number)
    SELECT SUBSTRING_INDEX(SUBSTRING_INDEX('a,b,c', ',', 2), ',', -1);
    

    有了这样的逻辑,为了在我们的列表中获得第3个标记,我们使用:

    SELECT SUBSTRING_INDEX(SUBSTRING_INDEX(GROUP_CONCAT(tags SEPARATOR ','), ',', 3), ',', -1)
    FROM test;
    

    返回:

    +-------------------------------------------------------------------------------------+
    | SUBSTRING_INDEX(SUBSTRING_INDEX(GROUP_CONCAT(tags SEPARATOR ','), ',', 3), ',', -1) |
    +-------------------------------------------------------------------------------------+
    | fried-rice                                                                          |
    +-------------------------------------------------------------------------------------+
    1 row in set (0.00 sec)
    

    将所有值作为不同的行

    我的想法有点棘手:

    1. 我知道我们可以通过连接表来创建行
    2. 我需要使用上面的请求
    3. 获取列表中的第N个标记

      因此,我们将创建一个表,其中包含从1到您列表中可能包含的最大标记数的所有数字。如果您可以拥有1M值,请创建1到1,000,000的1M条目。对于100个标签,这将是:

      CREATE TABLE numbers (
        num INT PRIMARY KEY
      );
      
      INSERT INTO numbers VALUES
          ( 1 ), ( 2 ), ( 3 ), ( 4 ), ( 5 ), ( 6 ), ( 7 ), ( 8 ), ( 9 ), ( 10 ), 
          ( 11 ), ( 12 ), ( 13 ), ( 14 ), ( 15 ), ( 16 ), ( 17 ), ( 18 ), ( 19 ), ( 20 ), 
          ( 21 ), ( 22 ), ( 23 ), ( 24 ), ( 25 ), ( 26 ), ( 27 ), ( 28 ), ( 29 ), ( 30 ), 
          ( 31 ), ( 32 ), ( 33 ), ( 34 ), ( 35 ), ( 36 ), ( 37 ), ( 38 ), ( 39 ), ( 40 ), 
          ( 41 ), ( 42 ), ( 43 ), ( 44 ), ( 45 ), ( 46 ), ( 47 ), ( 48 ), ( 49 ), ( 50 ), 
          ( 51 ), ( 52 ), ( 53 ), ( 54 ), ( 55 ), ( 56 ), ( 57 ), ( 58 ), ( 59 ), ( 60 ), 
          ( 61 ), ( 62 ), ( 63 ), ( 64 ), ( 65 ), ( 66 ), ( 67 ), ( 68 ), ( 69 ), ( 70 ), 
          ( 71 ), ( 72 ), ( 73 ), ( 74 ), ( 75 ), ( 76 ), ( 77 ), ( 78 ), ( 79 ), ( 80 ), 
          ( 81 ), ( 82 ), ( 83 ), ( 84 ), ( 85 ), ( 86 ), ( 87 ), ( 88 ), ( 89 ), ( 90 ), 
          ( 91 ), ( 92 ), ( 93 ), ( 94 ), ( 95 ), ( 96 ), ( 97 ), ( 98 ), ( 99 ), ( 100 );
      

      现在,我们使用以下查询得到num {num是number中的一行):

      SELECT n.num, SUBSTRING_INDEX(SUBSTRING_INDEX(all_tags, ',', num), ',', -1) as one_tag
      FROM (
        SELECT
          GROUP_CONCAT(tags SEPARATOR ',') AS all_tags,
          LENGTH(GROUP_CONCAT(tags SEPARATOR ',')) - LENGTH(REPLACE(GROUP_CONCAT(tags SEPARATOR ','), ',', '')) + 1 AS count_tags
        FROM test
      ) t
      JOIN numbers n
      ON n.num <= t.count_tags
      

      返回:

      +-----+---------------------+
      | num | one_tag             |
      +-----+---------------------+
      |   1 | pho                 |
      |   2 | pork                |
      |   3 | fried-rice          |
      |   4 | chicken             |
      |   5 | fried-rice          |
      |   6 | pork                |
      |   7 | chicken-calzone     |
      |   8 | chicken             |
      |   9 | fettuccine          |
      |  10 | chicken             |
      |  11 | spaghetti           |
      |  12 | chicken             |
      |  13 | spaghetti           |
      |  14 | chorizo             |
      |  15 | spaghetti           |
      |  16 | meat-balls          |
      |  17 | miso-soup           |
      |  18 | chanko-nabe         |
      |  19 | chicken-manchurian  |
      |  20 | chicken             |
      |  21 | manchurain          |
      |  22 | pork-manchurian     |
      |  23 | pork                |
      |  24 | manchurain          |
      |  25 | sweet-and-sour-pork |
      |  26 | pork                |
      |  27 | peking-duck         |
      |  28 | duck                |
      +-----+---------------------+
      28 rows in set (0.01 sec)
      

      计数标签出现

      只要我们现在有经典行,我们就可以轻松计算每个标记的出现次数。

      请参阅top of this answer以查看请求。

答案 1 :(得分:6)

Alain Tiembo有一个很好的答案,解释了很多机制。但是,他的解决方案需要一个临时表(数字)来解决问题。作为后续回答,我将他的所有步骤合并为一个单独的查询(使用tablename作为原始表格):

    SELECT t.tags, count(*) AS occurence FROM
    (SELECT
      tablename.id,
      SUBSTRING_INDEX(SUBSTRING_INDEX(tablename.tags, ',', numbers.n), ',', -1) tags
    FROM
      (SELECT 1 n UNION ALL SELECT 2
       UNION ALL SELECT 3 UNION ALL SELECT 4) numbers INNER JOIN tablename
      ON CHAR_LENGTH(tablename.tags)
         -CHAR_LENGTH(REPLACE(tablename.tags, ',', ''))>=numbers.n-1
    ORDER BY
      id, n) t
    GROUP BY t.tags
    ORDER BY occurence DESC, t.tags ASC

出于演示目的,请参阅 SQLFiddle

答案 2 :(得分:2)

首先,您应该使用联结表存储它,每个帖子和标记一行。但是,有时我们无法控制我们正在使用的数据结构。

假设您有一个有效标签列表,您可以执行您想要的操作:

select vt.tag, count(t.postid) as cnt
from validtags vt left join
     table t
     on find_in_set(vt.tag, t.tags) > 0
group by vt.tag
order by cnt desc;

答案 3 :(得分:1)

建议的方法是不在单个列中存储多个值,而是创建一个交集表。

因此,您的表格将包含以下列:
 1.标签:tag_id,名称
 2.帖子:post_id,category_code
 3. int_tags_to_posts:post_id,tag_id

获取计数:
select t.name, count(*) from tags t, posts p, int_tags_to_posts i where i.post_id = p.post_id and i.tag_id = t.tag_id group by i.tag_id order by count(*) desc;

答案 4 :(得分:0)

这应该有效:

SELECT tag, count(0) count FROM (
    SELECT tOut.*, REPLACE(SUBSTRING(SUBSTRING_INDEX(tags, ',', ocur_rank), LENGTH(SUBSTRING_INDEX(tags, ',', ocur_rank - 1)) + 1), ',', '') tag
      FROM (
        SELECT @num_type := if(@id_check = tY.id, @num_type + 1, 1) AS ocur_rank, @id_check := tY.id as id_check, tY.*
          FROM (
            SELECT LENGTH(tags) - LENGTH(REPLACE(tags, ',', '')) AS num_ocur, id, tags FROM tablename
          ) tX
          INNER JOIN (SELECT LENGTH(tags) - LENGTH(REPLACE(tags, ',', '')) AS num_ocur, id, tags FROM tablename) tY
          INNER JOIN (SELECT @num_type := 0, @id_check := 'some_id') tZ
       ) tOut
     WHERE ocur_rank <= num_ocur + 1
) tempTable GROUP BY tag ORDER BY count DESC;

将“tablename”替换为您的表名。

这个答案来自Jesse Perring在本页上发布的解决方案:

http://dev.mysql.com/doc/refman/5.0/en/string-functions.html#c12113