Writing a SQL conditional statement across columns for survey data - sql

Question

I am working with survey data that is feeding directly into a Snowflake DB.
In the survey tool, the primary question is formatted with a scaled numeric 0 through 10 satisfaction question, and then a subsequent question asks for a written explanation/exposition, but that subquestion can be skipped/left unanswered.
Also, based on how the respondent responds and rates to the scaled question, the survey tool will change the subquestion ie What do you like most about our Waffles? or What could be better?
Each of the three unique questions -the primary scaled question and the two different subquestions- have their own unique QUESTIONID and survey participants have a unique ID as RESPONSEID.
I am attempting to collect survey responses from only participants that have answered both the primary question and their resulting subquestion regardless of which subquestion they were prompted, and ignoring participants that did not write/respond to their subquestions.
The current query and output looks like this:
SELECT DISTINCT
r.RESPONSEID,
r.QUESTIONID,
CASE
WHEN r.SURVEY = q.SURVEY AND r.questionid = q.QUESTIONID THEN q.QUESTIONTEXT
ELSE NULL
END QTEXT,
r.RESPONSE,
FROM RESPONSES r
JOIN QUESTIONS q ON q.questionid = r.questionid
JOIN QUESTION_RESPONSE s ON s.response_id = r.responseid
WHERE r.SURVEY IN ('WaffleSurvey3000')
AND (q.QUESTIONID = '1' OR q.QUESTIONID = '1A'
OR q.QUESTIONID = '1B')
AND QTEXT IS NOT NULL
ORDER BY RESPONSEID;
And output:
RESPONSEID QUESTIONID QTEXT RESPONSE
A 1 Between 0 and 10... 7
B 1 Between 0 and 10... 9
B 1A What did you like... Best Waffles EVER!
C 1 Between 0 and 10... 5
D 1 Between 0 and 10... 6
E 1 Between 0 and 10... 2
E 1B What could be better... Awful Waffles! Do better! SHAME
But would like an output that ignores participants that did not answer their prompted subquestion asking for text which would look like:
RESPONSEID QUESTIONID QTEXT RESPONSE
B 1 Between 0 and 10... 9
B 1A What did you like... Best Waffles EVER!
E 1 Between 0 and 10... 2
E 1B What could be better... Awful Waffles! Do better! SHAME
I was hoping I could pull this off with another WHERE clause at the end, but since this seems to be a conditional statement, I'm assuming it'll involve another CASE statement, but not sure.
Any thoughts?

Answer 1

Perhaps you can add to the where clause of your query another filter:
AND r.RESPONSEID = (
select p.RESPONSEID from RESPONSES p
group by p.RESPONSEID having COUNT(1) > 1
)
Another version:
AND r.RESPONSEID||SUBSTR(r.QUESTIONID,1,1) = (
select p.RESPONSEID||SUBSTR(p.QUESTIONID,1,1) from RESPONSES p
group by p.RESPONSEID||SUBSTR(p.QUESTIONID,1,1) having COUNT(1) > 1
)