SQL¶

In [ ]:

! pip install sqlalchemy

In [ ]:

from kgforge.core import KnowledgeGraphForge

In [ ]:

forge = KnowledgeGraphForge("../../configurations/forge.yml")

Imports¶

In [ ]:

from kgforge.core import Resource

In [ ]:

import pandas as pd

In [ ]:

from sqlalchemy import create_engine

In [ ]:

engine = create_engine("sqlite:///:memory:")

In [ ]:

jane = Resource(name="Jane Doe", email="jane.doe@epfl.ch")

In [ ]:

john = Resource(name="John Smith", email="john.smith@epfl.ch")

In [ ]:

other = Resource(name="John Doe", email="john.doe@unige.ch")

In [ ]:

persons = [jane, john, other]

Converting Resources to DataFrame

In [ ]:

df = forge.as_dataframe(persons)

Persisting Resources in the SQL Database.

NB: The more powerful way is to have a Store implementation for the SQL Database.

In [ ]:

df.to_sql("Persons", engine, index=False)

Getting the table as a DataFrame

In [ ]:

pd.read_sql("SELECT * FROM Persons", engine)

Filtering on the SQL Database to keep only persons with an EPFL email

In [ ]:

df = pd.read_sql("SELECT * FROM Persons WHERE email LIKE '%@epfl.ch'", engine)

In [ ]:

df

Converting DataFrame to Resources

In [ ]:

persons = forge.from_dataframe(df)

In [ ]:

print(*persons, sep="\n")

In [ ]: