https://kedro.org/ logo
#questions
Title
# questions
g

Gabriel Bandeira

09/06/2023, 9:06 PM
[SOLVED] Hi, team! I’m trying to load a few files stored on GCP with my kedro pipeline but I’m getting the error
Copy code
DatasetError: 
'refresh_token'.
It happens on all files I try to load (
xlsx
,
pkl
,
parquet
). Any clues on that?
Copy code
Python 3.10.12
kedro==0.18.12
kedro-datasets[spark,pandas]==1.5.3
gcsfs==2023.9.0
On my credentials file I don’t have the key
refresh_token
I only have the keys:
Copy code
type
project_id
private_key_id
private_key
client_email
client_id
auth_uri
token_uri
auth_provider_x509_cert_url
client_x509_cert_url
universe_domain
I am able to connect to the GCP account using this credentials for loading and saving directly on BigQuery but I need to open a link and authenticate on the browser to do it.
[SOLUTION] I needed to authenticate using
GOOGLE_APPLICATION_CREDENTIALS
environment variable pointing to a json with credentials
🚀 2
e

Elaine Resende

10/26/2023, 9:55 PM
Hey Gabriel, how is you credentials file? Which keys it has? I have generated mine but i am getting a DatasetError
g

Gabriel Bandeira

11/16/2023, 5:49 PM
Hi, @Elaine Resende! Sorry for the late response, but I just saw it now. My credentials file is a json with the following keys: 1. type 2. project_id 3. private_key_id 4. private_key 5. client_email 6. client_id 7. auth_uri 8. token_uri 9. auth_provider_x509_cert_url 10. client_x509_cert_url 11. universe_domain
e

Elaine Resende

11/20/2023, 7:52 PM
Thanks