How to Correctly Sum Aggregate Values from Joined Tables in MySQL to Avoid Cross Product Issues?

Using SUM aggregation to join tables in MySQL: solving cross-product problems

Combining two queries using the SUM aggregate function often presents challenges in MySQL. Cross products between tables can cause incorrect SUM values. To solve this problem, the SUM function needs to be encapsulated into a subquery.

Consider the following example where two queries retrieve SUM(drive_time) and SUM(tm_hours) respectively for a specific date and a teacher ID of 5:

Query 1:

SELECT last_name, first_name, DATE_FORMAT(mil_date, '%m/%d/%y') AS dates, 
SUM(drive_time) MINUTES 
FROM bhds_mileage 
LEFT JOIN bhds_teachers i 
ON i.ds_id = bhds_mileage.ds_id 
WHERE mil_date BETWEEN '2016-04-11' AND '2016-04-30' 
AND bhds_mileage.ds_id = 5 
GROUP BY CONCAT(YEAR(mil_date), '/', WEEK(mil_date)), bhds_mileage.ds_id 
ORDER BY last_name ASC, dates ASC

Query 2:

SELECT last_name, first_name, DATE_FORMAT(tm_date, '%m/%d/%y') AS dates,   
SUM(tm_hours) total 
FROM bhds_timecard 
LEFT JOIN bhds_teachers i 
ON i.ds_id = bhds_timecard.ds_id 
WHERE tm_date BETWEEN '2016-04-11' AND '2016-04-30' AND bhds_timecard.ds_id = 5 
GROUP BY CONCAT(YEAR(tm_date), '/', WEEK(tm_date)), bhds_timecard.ds_id 
ORDER BY last_name ASC, dates ASC

Simple connection attempt:

To combine these queries, a simple way is to concatenate them as follows:

SELECT last_name, first_name, DATE_FORMAT(tm_date, '%m/%d/%y') AS dates,  
SUM(tm_hours) total, SUM(drive_time) MINUTES 
FROM bhds_timecard 
LEFT JOIN bhds_teachers i 
ON i.ds_id = bhds_timecard.ds_id 
LEFT JOIN bhds_mileage 
ON DATE_FORMAT(bhds_timecard.tm_date, '%m/%d/%y') = DATE_FORMAT(bhds_mileage.mil_date, '%m/%d/%y') AND bhds_timecard.ds_id = bhds_mileage.ds_id 
WHERE tm_date BETWEEN '2016-04-11' AND '2016-04-30' AND bhds_timecard.ds_id = 5 
GROUP BY CONCAT(YEAR(tm_date), '/', WEEK(tm_date)), bhds_timecard.ds_id 

However, this approach creates cross products between tables, resulting in incorrect SUM values.

Solution:

To get correct SUM values, move the SUM aggregate function into a subquery. This prevents cross-product problems and ensures that SUM values ​​are only calculated in relevant rows:

SELECT last_name, first_name, DATE_FORMAT(LEAST(mil_date, tm_date), '%m/%d/%y') AS dates, 
        total, minutes 
FROM bhds_teachers AS i 
LEFT JOIN ( 
    SELECT ds_id, YEARWEEK(mil_date) AS week, MIN(mil_date) AS mil_date, SUM(drive_time) AS minutes 
    FROM bhds_mileage 
    WHERE mil_date BETWEEN '2016-04-11' AND '2016-04-30' 
    AND bhds_mileage.ds_id = 5 
    GROUP BY ds_id, week 
) AS m 
ON m.ds_id = i.ds_id 
LEFT JOIN ( 
    SELECT ds_id, YEARWEEK(tm_date) AS week, MIN(tm_date) AS tm_date, SUM(tm_hours) AS total 
    FROM bhds_timecard
    WHERE tm_date BETWEEN '2016-04-11' AND '2016-04-30' AND bhds_timecard.ds_id = 5 
    GROUP BY ds_id, week 
) AS t 
ON t.ds_id = i.ds_id AND t.week = m.week 

The above is the detailed content of How to Correctly Sum Aggregate Values from Joined Tables in MySQL to Avoid Cross Product Issues?. For more information, please follow other related articles on the PHP Chinese website!