Maplib dealing with missing values from input CSV

[henrieglesorotos]

Looking at the examples - http://github.com/DataTreehouse/maplib/blob/main/examples/mapping/mapping.ipynb, I cannot see any that show missing values. Is this possible? Eg:

orgIRI,orgName,numEmployees,foundingDate,description
http://test.com/company/acme,Acme Corp,100,,Technology company
http://test.com/company/techco,TechCo Inc,50,2019-06-01,Software development

See above CSV where we don't know the foundingDate of Acme Corp. Will this throw an error? Does this necessitate and different template for each row if using programmatic templates?

[henrieglesorotos]

is this something that is possible @magbak ?

[magbak]

Hi @henrieglesorotos,
Sorry about the delayed reply and thank you for the reminder.
OTTR has good support for missing data.

Here is some code:

import polars as pl
from io import StringIO
from maplib import Model

csv = """orgIRI,orgName,numEmployees,foundingDate,description
http://test.com/company/acme,Acme Corp,100,,Technology company
http://test.com/company/techco,TechCo Inc,50,2019-06-01,Software development"""
csv_sio = StringIO(csv)

df = pl.read_csv(csv_sio)

m = Model()
template = """
@prefix ex: <urn:henri:example:> .
@prefix xsd:  <http://www.w3.org/2001/XMLSchema#> .
@prefix ottr: <http://ns.ottr.xyz/0.4/> .

ex:my_template [
     ottr:IRI ?orgIRI, 
     xsd:string ?orgName, 
     xsd:long ?numEmployees, 
    # We add a "?" before the datatype to make the parameter optional.
    ? xsd:string ?foundingDate, 
     xsd:string ?description ] :: {
  ottr:Triple(?orgIRI, ex:orgName, ?orgName) ,
  ottr:Triple(?orgIRI, ex:numEmployees, ?numEmployees) ,
  ottr:Triple(?orgIRI, ex:foundingDate, ?foundingDate) ,
  ottr:Triple(?orgIRI, ex:description, ?description)
} . """

m.add_template(template)
m.map("ex:my_template", df)

m.query("""SELECT * WHERE {
?a ?b ?c
}""")

This produces the following result:

[henrieglesorotos]

oh you legend @magbak ! Is this functionality exposed using the python implementation of the templates?

[magbak]

oh you legend @magbak ! Is this functionality exposed using the python implementation of the templates?

Yes! Just use the optional=True flag in the Parameter class.
See: https://datatreehouse.github.io/maplib/maplib.html#Parameter