SQL | Parent Child relationship in same table - postgresql

Question

I have a table that has parent and child key relationship in the same table, I need to find Parent rows that don't have any children(for example row 1 have no other children) and the most recent children (for example like row 17 have 3 children i.e. 10,13,14 and we need to fetch most recent children only which is 10 )

Answer 1

You need one part related to be main parent:
SELECT parent.*
FROM MyTable AS parent
WHERE parent_id = 0
Next you need to find most recent direct child:
SELECT parent.*, child.*
FROM MyTable AS parent
LEFT JOIN MyTable AS child
ON child.parent_id = parent.id
WHERE parent_id = 0
AND RANK() OVER (PARTITION BY child.parent_id ORDER BY child.id DESC) = 1
And add no-child ones:
SELECT parent.*, child.*
FROM MyTable AS parent
LEFT JOIN MyTable AS child
ON child.parent_id = parent.id
WHERE parent_id = 0
AND (
RANK() OVER (PARTITION BY child.parent_id ORDER BY child.id DESC) = 1
OR child.id IS NULL)

Answer 2

Something like this should work, but may not be the most efficient way of doing it:
select {desired fields}
from your_table
where not exists
(
select 1 from your_table B
where B.parent_id = your_table.id
)
UNION
select {desired fields}
from your_table
where id in
(
select max(id)
from your_table B
where B.parent_id = your_table.id
)

Answer 3

Hmmm . . . You seem to want one row per parent. Either the most recent child or the parent if there are no children. You can actually do this with window functions:
select t.*
from (select t.*,
row_number() over (partition by coalesce(nullif(parent_id, 0), id
order by parent_id desc, id desc
) as seqnum
from t
) t
where seqnum = 1;