Question

I’m trying to use the SUM function to count rows from 3 tables, which is however, not working effectively since when the total_files and total_notes are returned, they both are the same when there is at least one file and then total_files will take the same value as total_notes which I don’t understand why it’s doing that.

It should count the number of rows which is relevant to each record that will get return as a record list with a count of total files, total notes and total contacts assigned to the record per record row (the data of files, notes and contacts do not get displayed only counted).

My query is shown below:

SELECT rec.street_number,
       rec.street_name,
       rec.city,
       rec.state,
       rec.country,
       rec.latitude,
       rec.longitude,
       LEFT(rec.description, 250) AS description,
       usr.username,
       usr.full_name,
       ppl.person_id,
       ppl.first_name,
       ppl.last_name,
       SUM(IF(rlk.record_id = rec.record_id, 1, 0)) AS total_contacts,
       SUM(IF(files.record_id = rec.record_id, 1, 0)) AS total_files,
       SUM(IF(notes.record_id = rec.record_id, 1, 0)) AS total_notes,
       (
           SELECT COUNT(DISTINCT rec.record_id)
           FROM records rec
           WHERE rec.marked_delete = 0 AND rec.is_archive = 0
       ) AS total_records
FROM
(
    records rec

    INNER JOIN members usr ON rec.user_id = usr.user_id

    LEFT OUTER JOIN record_links rlk ON rec.record_id = rlk.record_id

    LEFT OUTER JOIN people ppl ON ppl.person_id = rlk.person_id AND rlk.record_id = rec.record_id

    LEFT OUTER JOIN files files ON files.record_id = rec.record_id

    LEFT OUTER JOIN notes notes ON notes.record_id = rec.record_id
)
WHERE rec.marked_delete = 0 AND rec.is_archive = 0
GROUP BY rec.record_id
ORDER BY rec.submit_date DESC
LIMIT 0, 25

Basically as you can see there is three SUM which will count relevant rows that comes from those tables, but I seriously don’t understand how total_files would be taking the same value as total_notes is there something wrong I’m doing here?

Answer 1

It’s because rec is joined to both notes and files.

Suppose record 1 has 2 notes and 1 file, record 2 has two note and two files, and record 3 has a note but no files.

Then the table rec LEFT OUTER JOIN files ... LEFT OUTER JOIN notes will look like this:

+-----------+---------+---------+
| record_id | file_id | note_id |
+-----------+---------+---------+
|         1 |       1 |       1 |
|         1 |       2 |       1 |
|         2 |       3 |       2 |
|         2 |       3 |       3 |
|         2 |       4 |       2 |
|         2 |       4 |       3 |
|         3 |    NULL |       4 |
+-----------+---------+---------+

Note how every file_id gets joined to every note_id (within the same record_id). Also, since you have SUM(IF(files.record_id = rec.record_id,1,0)) and the join condition is files.record_id = rec.record_id, you are actually counting COUNT(files)*COUNT(notes) per record_id.

I’d recommend you instead COUNT(DISTINCT files.id) and COUNT(DISTINCT records.id). The column in the COUNT would be your primary key on files/notes, not files.record_id:

SELECT rec.record_id,
       COUNT(DISTINCT files.id) AS total_files,
       COUNT(DISTINCT notes.id) AS total_notes
FROM rec
-- note: LEFT OUTER JOIN is the same as LEFT JOIN in MySQL
LEFT JOIN files ON files.record_id=rec.record_id 
LEFT JOIN notes ON notes.record_id=rec.record_id
GROUP BY record_id


+-----------+-------------+-------------+
| record_id | total_files | total_notes |
+-----------+-------------+-------------+
|         1 |           2 |           1 |
|         2 |           2 |           2 |
|         3 |           0 |           1 |
+-----------+-------------+-------------+

Of course, adjust to your query as necessary (add in those extra columns/joins).