alteryx / alteryx/featuretools

Ability to infer relationships between two df's

Aperta
#866 0 commenti 0 reazioni 0 assegnatari Vedi su GitHub
API needs design
Lingua principale
Python
Stelle
7.7k
Fork
915
Metriche di merge delle PR
Nessuna PR unita negli ultimi 30g

Descrizione

It can be useful to be able to infer relationships between two tables. Especially as we build higher level applications on top of featuretools.

This functionality could be implemented based on rules and heuristics

Rules:
* relationships must be between two variables of same dtype
* the parent variable must be the index column

Heuristics:
* relationships often taken the form of `product_id --> id` or `product_id` --> `product_id-->`
* the child id shouldn't be a numeric semantic type (it's fine if the underlying data is numeric)

```python

def recommend_relationships(entity_a, entity_b):
"""Returns potential relationships between entity a and entity b

Args:
entity_a (ft.Entity): entity a
entity_b (ft.Entity): entity b

Returns:
List[ft.Relationship]: list of potential relationships
"""
pass
```

it may also also make sense to do an api that takes in a full entityset

```python

def recommend_relationships_entityset(entityset):
"""Returns all potential relationships between two entities in an entityset

Args:
entityset (ft.EntitySet): entityset

Returns:
List[ft.Relationship]: list of potential relationships
"""
pass
```

Guida per i contributori

Apri la guida per i contributori

Valutazione

Questa issue non è ancora stata valutata.

Ricevi le nuove issue nella tua casella

Un breve riepilogo di issue GitHub adatte ai principianti.