[Question] Passing Pandas DataFrames as input

[AchilleSalaun]

Hi,

From my understanding, it is possible to load directly SPARQL-Anything's outputs into rdflib graphs or Pandas DataFrames, but is it possible to apply SPARQL-Anything to Python variables (e.g. a Pandas DataFrame) or does the tool limit to inputing files?

[MarcoR1791]

Hi,

Yes, your understanding is correct. At the moment the tool can output Python objects of selected types but cannot directly receive those as inputs (See #2 for example).

What's your specific use-case?

[AchilleSalaun]

Hi,

Thanks for your quick answer.

I am working on the design of a Python-based ETL pipeline that returns triples from various tabular data files. I would like to anticipate the possibility of preprocessing data using Pandas before triplification using SPARQL-Anything. As it is, I see the need of saving intermediate files just to input data into SPARQL-Anything and I was wondering about the existence of a smoother process?

You're correct, my question is indeed the third point of the issue #2 where Pandas DataFrames add up to the list of the mentioned Python objects. My apologies, I completely missed that point while writing this issue.