Is it possible the the move to `StructuredDataset`...
# ask-the-community
s
Is it possible the the move to
StructuredDataset
broke support for schema urls in the flyteconsole launch form? Not sure if it ever worked, but https://github.com/flyteorg/flyte/issues/405 seems to indicate it did. Given this example
Copy code
@task
def t1(df: pd.DataFrame) -> pd.DataFrame:
    return df

@workflow
def wf(df: pd.DataFrame) -> pd.DataFrame:
    return t1(df=df)
I get the following message in the launch form:
y
@Jason Porter was this added? i thought it was
s
Yeah. There's a problem with launching workflows that accept structured dataset as an input. We'll need to fix this. Despite giving a default value, it still isn't accepting the structured dataset and resulting in an error.
j
Noted: we'll pick it up next spring đź‘Ť https://github.com/flyteorg/flyte/issues/3395
s
It looks like this is also a restriction in flytectl when generating execFiles:
Copy code
flytectl get launchplan my_launchplan --execFile exec.yaml

error creating default value for literal type  structured_dataset_type:<format:"parquet" > 
Error: failed to convert to a known Literal. Input Type [structured_dataset_type:<format:"parquet" > ] not supported
Should I create an issue?
s
@Prafulla Mahindrakar, you did some structured datasets work in flytectl recently, correct? Should this still be handled in flytectl?
p
yes this should have been working with latest flytectl . Could some one help with getting me example workflow or reference from cookbook which i can register and try creating an exec spec from it using flytectl .
p
yeah seems we need more changes for these. I can track improvements in flytectl as part of the same FE issue.
100 Views