:wave: I'm getting the error `AvroException: ('Dat...
# troubleshootc
calm-balloon-31412
02/03/2023, 4:49 PM👋 I'm getting the error
AvroException: ('Datum union type not in schema: %s', None)graph.get_aspect_v2(entity_urn=urn, aspect="dataJobInfo", aspect_type=DataJobInfoClass)✅ 1
b
bulky-soccer-26729
02/03/2023, 5:24 PMhey Divya! are you trying to access the aspect in the same way as the person you linked to? as in you're trying to use the python SDK and it's not working?
c
calm-balloon-31412
02/03/2023, 5:27 PMI'm trying to access the aspect from an Airflow job, doing something like this for the setup:
Copy code
gms_endpoint = "<http://datahub-gms.service.envoy:10080>"
graph = DataHubGraph(DatahubClientConfig(server=gms_endpoint))
urn = "..."
info = graph.get_aspect_v2(entity_urn=urn, aspect="dataJobInfo", aspect_type=DataJobInfoClass)b
bulky-soccer-26729
02/03/2023, 5:34 PMare you having issues with getting aspects like this with any other aspect type? also tbh I'm not the expert on our python SDK, but can you try getting it without the
aspect="dataJobInfo"info = graph.get_aspect_v2(entity_urn=urn, aspect_type=DataJobInfoClass)c
calm-balloon-31412
02/03/2023, 5:35 PMThe issue is only with this specific aspect, I'm able to retrieve the other ones just fine (such as glossary terms)
Copy code
get_aspect_v2() missing 1 required positional argument: 'aspect'b
bulky-soccer-26729
02/03/2023, 5:36 PMahh it's the v2 version sorry I didn't actually realize that
bulky-soccer-26729
02/03/2023, 5:37 PMhmm I'm going to try this myself and see if I can get any extra help here
c
calm-balloon-31412
02/03/2023, 5:37 PMI get the same error with
get_aspect()calm-balloon-31412
02/03/2023, 5:37 PMthank you!
b
bulky-soccer-26729
02/03/2023, 5:49 PMokay good to know about
get_aspect()m
mammoth-bear-12532
02/03/2023, 10:00 PM
@calm-balloon-31412: try this
Copy code
from datahub.metadata.schema_classes import DataJobInfoClass
from datahub.ingestion.graph.client import DataHubGraph, DataHubGraphConfig
graph = DataHubGraph(DataHubGraphConfig())
datajob = graph.get_aspect(entity_urn="urn:li:dataJob:(urn:li:dataFlow:(airflow,dag_abc,PROD),task_123)", aspect_type=DataJobInfoClass)
print(datajob)mammoth-bear-12532
02/03/2023, 10:03 PM
if you're still facing this error, there is likely a mismatch between the cli version (model) and the server version (model)
c
calm-balloon-31412
02/03/2023, 10:03 PMMaybe because we are on an older version, but I don't have
DataHubGraphConfigdatahub.ingestion.graph.clientm
mammoth-bear-12532
02/03/2023, 10:05 PM
yeah that shouldn't be causing a problem
mammoth-bear-12532
02/03/2023, 10:06 PM
the main issue is probably your client is compiled against a version of the model that doesn't have an enum value ... while the server is sending it back
mammoth-bear-12532
02/03/2023, 10:06 PM
upgrading the client should solve this
mammoth-bear-12532
02/03/2023, 10:06 PM
another workaround would be to the
cli_utilsmammoth-bear-12532
02/03/2023, 10:15 PM
here is what that would look like:
Copy code
from datahub.cli.cli_utils import get_aspects_for_entity
from datahub.cli.cli_utils import ENV_METADATA_HOST_URL
import os
def set_host_url(host_url: str):
os.environ.set(ENV_METADATA_HOST_URL, host_url)
set_host_url("<http://localhost:8080>")
aspect_dict = get_aspects_for_entity(entity_urn="urn:li:dataJob:(urn:li:dataFlow:(airflow,dag_abc,PROD),task_123)", aspects=["dataJobInfo"], typed=False)
print(aspect_dict)mammoth-bear-12532
02/03/2023, 10:15 PM
you would get an untyped dictionary back
c
calm-balloon-31412
02/03/2023, 10:18 PMoh cool that worked!!
calm-balloon-31412
02/03/2023, 10:18 PMthanks so much 🙂
5 Views