Question

I am creating a query below where a it retrieves data depending on the term(s) the user entered in the a search bar. Now what I am trying to do is not display duplicate data, so if there are 2 rows where all the fields are exactly the same, then it is a duplicate row, so it only shows this row once not multiple times. Now I think it seems to do this as I see no duplicate rows but all I did was do a GROUP BY with all the SELECT fields except for the Answer field as it doesn’t let me have a group concat in the GROUP BY clause.

But what my question is that do I need that field in the GROUP Y clause to not show duplicate rows or is it not really needed?

SELECT
      q.QuestionContent, 
      o.OptionType, 
      q.NoofAnswers, 
      GROUP_CONCAT(DISTINCT Answer ORDER BY Answer SEPARATOR ',') AS Answer, 
      r.ReplyType, 
      q.QuestionMarks
    FROM Question q
    LEFT JOIN Answer an
      ON q.QuestionId = an.QuestionId
    LEFT JOIN Reply r 
      ON q.ReplyId = r.ReplyId 
    LEFT JOIN Option_Table o 
      ON q.OptionId = o.OptionId 
    WHERE ".implode(" AND ", array_fill(0, $numTerms, "q.QuestionContent LIKE ?"))."
    GROUP BY   q.QuestionContent, 
      o.OptionType, 
      q.NoofAnswers, 
      r.ReplyType, 
      q.QuestionMarks
    ORDER BY ".implode(", ", array_fill(0, $numTerms, "IF(q.QuestionContent LIKE ?, 1, 0) DESC"))."

Answer 1

No, you do not need the group_concat() in the select. The group by will ensure that any particular combination of values will appear once for the columns in the group by. Because these are guaranteed to be distinct, you don’t have to worry about any other columns.

The group_concat() is a calculated column, based on the summaries. You are not permitted to have such columns in a group by statement. If you want to aggregate on them again, then you need to use a subquery.