It works not only for DATE attributes but for all attributes

hii please reply

@Jan Soubusta please reply

There is no GROUP BY in your query. In the SELECT section you use only UCD.duration. The LEFT OUTER JOIN does not change anything here. If you would SELECT SUM() without JOIN, the result would be the same. Here is the query where the OUTER JOIN plays active role:

SELECT
	SUM(UCD.duration)/ 3600 AS hrs,
	UCA.course_assignment_name
FROM
	user_course_assignments AS UCA
LEFT OUTER JOIN user_course_durations AS UCD 
	ON UCD.user_course_id = UCA.course_id
GROUP BY
	UCA.course_assignment_name
;

In this case, I want to select all course assignments (their names) and their SUM(duration), if a corresponding record in

user_course_durations

exists, otherwise NULL. Additionally, you could use COALESCE function and define default when NULL, e.g. 0. This would be managed in MAQL like this:

SELECT SUM({fact/duration}) FOR EACH {label/course_assignment_name}

Or more general solution is to specify the whole dataset after FOR EACH:

SELECT SUM({fact/duration}) FOR EACH {dataset/course_assignments}

Then, when you put such a metric to a visualization and add any attribute/label from

user_course_assignments

, we generate OUTER JOIN under-the-hood.