SELECT column different from GROUP BY column in hierarchy structure - mysql

Question

I am using SQL to store information about a family tree. I want to be able to input the name of a person born in Generation X and output the names of all of their descendants born in Generation Z, but only if these descendants are brother and sister (have the same parent). I have found similar solutions, but I have found these solutions do not work with a hierarchical structure (due to aliasing). There may also be a parent and their child born in generation Y, so hierarchy depth is not reliable to determine Names.
+----+----------+-----------+------------+--------+
| ID | Name | Parent_ID | Generation | Gender |
+----+----------+-----------+------------+--------+
| 1 | John | NULL | X | Male |
| 2 | Jill | 1 | Y | Female |
| 3 | Andy | 2 | Z | Male |
| 4 | Ralph | 2 | Z | Male |
| 5 | Lisa | NULL | X | Female |
| 6 | Steve | 5 | Y | Male |
| 7 | Sean | 6 | Y | Male |
| 8 | Sarah | 6 | Y | Female |
| 9 | Emily | 7 | Z | Female |
| 10 | Matt | 7 | Z | Male |
+----+----------+-----------+------------+--------+
Desired output: (if SET #GenX = 'Lisa';)
+-------+
| Name |
+-------+
| Emily |
| Matt |
+-------+
This is my code so far.
SET #GenX = 'Lisa';
SELECT t2.name AS Kids
FROM (SELECT name, Parent_ID
FROM FamilyTree
WHERE Gender = 'Male' OR Gender = 'Female') t1
LEFT JOIN FamilyTree t2 ON t2.ID = t1.Parent_ID
LEFT JOIN FamilyTree t3 ON t3.ID = t2.Parent_ID
LEFT JOIN FamilyTree t4 ON t4.ID = t3.Parent_ID
WHERE t3.name = '#GenX' OR t4.name - '#GenX'
GROUP BY t2.name
HAVING SUM(t1.Gender) > 0 AND SUM(t1.Gender) > 0;
This returns the correct parent name, but does not return the kids names. If I SELECT and GROUP BY the kids names, I can't find which kids have sisters/brothers. Thanks for the help. I'm really stuck on this one.
You may assume no repeated names or overlapping family trees.

Answer 1

;with cte
as
(
Select id,name,Parent_id,generation,gender from family_tree where name =#name and generation='x'
Union All
Select t.id, t.name,t.Parent_id,t.generation,t.gender from family_tree t
join cte c on c. id=t.parent_id
)
,condition
as
(
select Parent_id,COUNT(*) cnt from cte Where generation='z' Group by parent_id having COUNT(*)>1
)
,sibling
as
(
select parent_id,Count(gender) cntg from
(select distinct c.parent_id,c.gender from cte c join condition co on co.parent_id=c.parent_id) a
group by parent_id having count(gender)>1
)
select c.name from cte c join sibling s on c.parent_id=s.parent_id

Answer 2

Assuming that you have only 3 generation you can go with the following query
SELECT NAME FROM FamilyTree WHERE PARENT_ID IN
(SELECT ID FROM FamilyTree WHERE PARENT_ID IN
(SELECT ID FROM FamilyTree WHERE PARENT_ID =
(SELECT ID FROM FamilyTree WHERE NAME = 'Lisa')
)
) AND
(SELECT COUNT(ID) FROM
(SELECT ID FROM FamilyTree WHERE PARENT_ID IN
(SELECT ID FROM FamilyTree WHERE PARENT_ID IN
(SELECT ID FROM FamilyTree WHERE PARENT_ID =
(SELECT ID FROM FamilyTree WHERE NAME = 'Lisa')
)
)
)T2
)=2 AND GENARATION = 'Z'
Above query will give result if their is only 3 generation
If its more you have to add more inner query and it will effect you performance.

Answer 3

My original attempt returned the id's (names) of each parent of brother/sister Generation Z kids. The solution required using the IN operator to relate children's names to the parent id's given from the first nested SELECT statement.
SET #GenX = 'Lisa';
SELECT name FROM FamilyTree WHERE Parent_ID IN
(SELECT t2.id AS Kids
FROM
(SELECT name, Parent_ID
FROM FamilyTree
WHERE Gender = 'Male' OR Gender = 'Female') t1
-- Iterate through hierarchy
LEFT JOIN FamilyTree t2 ON t2.ID = t1.Parent_ID
LEFT JOIN FamilyTree t3 ON t3.ID = t2.Parent_ID
LEFT JOIN FamilyTree t4 ON t4.ID = t3.Parent_ID
WHERE t3.name = '#GenX' OR t4.name - '#GenX'
GROUP BY t2.id
HAVING SUM(t1.Gender) > 0 AND SUM(t1.Gender) > 0); -- includes only brother/sister siblings