如何制定查询以加入多个表

Question

MY tables are as follows

users

 1     Michael
 2     James
 3     John
 4     Susie
 5     Harvey

products

 1     learn_java        2           1         21       0             1
 2     learn_sql         1           2         8        0             2
 3     4 GB DDR3         0           3         5        0             3
 4     love jacket       2           4         0        0             5
 5     1 TB HDD          3           5         12       1             4
 6     kill_ants         3           6         5        0             6
 7     2 TB HDD          2           7         2        0             4
 8     8 GB DDR3         2           8         18       0             3
 9     1 GB DDR2         3           9         7        0             3

product_group

1      text    1
2      pdf     1
3      ram     2
4      hdd     2
5      leather 0
6      diy     0

product_category

1      book
2      electronics

/* forgot about comment field*/
comments

     1            1       ...
     2            1       ...
     3            2       ...
     4            2       ...
     5            2       ...
     6            3       ...

My Goal:

The product table has 4 types of data.

1. Products with no uploader( uploader = 0 )

2. Products that are in review( exclude = 1 )

3. Products that fall under category = 0(poducts.groupid = product_group.gid AND product_group.category = 0 )

4. Products that are uploaded by an uploader, not in review and not fall under category = 0( uploader != 0, exclude = 1, poducts.groupid = product_group.gid AND product_group.category != 0)

I only have to consider the 4th type of data. I have to exclude the first three types of data. I have to group these data by their uploader. Say, James have uploaded 3 product, Jones have uploaded 2 product and the rest of the user hasn't uploaded anything.

The query should return this

3 James SUM of views of 3 products 2 Jones SUM of views of 2 products 0 user1 0 0 user2 0 .... ....

So if I consider the data of my table, I want to get the data in follwing order

     3            2           James           41(21+18+2)
     1            3           Jones           7
     1            1           Michael         8
     0            5           Harvey          0/NULL
     0            4           Susie           0/NULL

I came up with this.

SELECT COUNT(pid) as product_num,
       SUM(views) as total_views
       users.*
FROM users
INNER JOIN products ON products.uploader = users.id
INNER JOIN product_group ON products.groupid = product_group.category
WHERE exclude = 0
AND product_group.category != 0

Which obviously doesn't work as it doesn't include the users, who hasn't uploaded any product. How to make this work to take these users into account?

EDIT:

SELECT COUNT(pid) as product_num,
       SUM(views) as total_views
       users.*
FROM users
LEFT JOIN products ON products.uploader = users.id
INNER JOIN product_group ON products.groupid = product_group.category
WHERE exclude = 0
AND product_group.category != 0
GROUP BY users.id
ORDER BY product_num

It also doesn't take users with 0 upload.

Second EDIT:

I have added a comment table(I forgot about it earlier). Is there any way to show the total_number of comments for a user.

Here, James has uploaded product 1, 4, 8. For these post_id is also 1, 4, 8(In real these won't be same). From comments table, these posts have following number of comments 2, 0, 0. So, total number of comment 2.

So, final result should be

     3            2           James           41(21+18+2)       2
     1            3           Jones           7                 0/NULL
     1            1           Michael         8                 3
     0            5           Harvey          0/NULL            0/NULL
     0            4           Susie           0/NULL            0/NULL

Answer 1

Try this query:

SELECT COUNT(pid) as product_num,
       SUM(views) as total_views,
       u.*
FROM users u
LEFT JOIN products p
  ON p.uploader = u.id
LEFT JOIN product_group pg
  ON p.groupid = pg.gid
WHERE p.exclude = 0 
  AND p.uploader <> 0 
  AND pg.category != 0
   OR p.pid is null
GROUP BY u.id

demo: http://sqlfiddle.com/#!2/c94bef/12

To count a number of comments, please add a dependent subquery to the SELECT clause:

         SELECT count(*)
         FROM comments c
         WHERE c.post_id = p.post_id

in this way:

SELECT COUNT(pid) as product_num,
       SUM(views) as total_views,
       ( SELECT count(*)
         FROM comments c
         WHERE c.post_id = p.post_id
       ) As total_comments,
       u.*
FROM users u
LEFT JOIN products p
  ON p.uploader = u.id
LEFT JOIN product_group pg
  ON p.groupid = pg.gid
WHERE p.exclude = 0 
  AND p.uploader <> 0 
  AND pg.category != 0
   OR p.pid is null
GROUP BY u.id

demo: http://sqlfiddle.com/#!2/5c44f/1