0
Something related to hierarchical data.
I am implementing a Labels system for messaging a website that is similar to the gmail Labels system. This is the short ER.
A message may have one or more Abels and each label may belong to more than one message.
My question is about the table tab_category
.
With that approach id -> id_father
i can have categories and subcategories virtually infinitely. Example:
id name father
1 catA 0
2 catAB 1
3 catAC 1
4 catABA 2
5 catACA 3
6 catB 0
7 catC 0
8 ... ...
This example can return:
- cata/catAB
- cata/catAC
- cata/catAB/catABA
- cata/catAC/Cataca
- catB
- catC
- ...
My question is whether this approach is appropriate for this type of scenario, thinking about performance, when we will have many categories and many sub-sub-categories? If not appropriate what is the modern approach to this scenario?
In case you are right to worry you would use a recursive query p/ find all decentines of a tag but recursiveness is limited in SQL and non-performance. One idea (I don’t know good or bad) is to use a relationship table that denotes the degree of kinship, e.g.: A => +2 => ABA, ABA => -2 =>A, A=>+1=>AB
– jean
Hi @jean vc could draw your idea? I was also thinking of an intermediate table to mitigate this problem.
– zwitterion