Hello, I hope that this finds you well! 'Partitio...
# questions
c
Hello, I hope that this finds you well! 'PartitionedDataset' not found I'm using a PartitionedDataset and have been (successfully) testing the config before but running
kedro run
throws the following error.
Copy code
DataSetError: Class 'PartitionedDataset' not found or one of its dependencies has not been installed.
my dataset config is the following
Copy code
# Raw data
optical_images_raw:
    type: PartitionedDataset
    path: data/01_raw/optical/
    dataset: my_pkg.extras.datasets.image_dataset.ImageDataSet
    filename_suffix: .tif
    overwrite: True
    layer: raw
and yes,
my_pkg...ImageDataset
is well installed. Help? Anyone has ever faced a similar issue? Google did not help much so far? Any fix idea? kedro version: 0.18.10
K 1
d
is there a chance there are any issues with the dependencies of the wrapped dataset?
c
ah! good question! let me check
d
my trick with these sorts of issues is to open a notebook and try and do this config with the python api
c
is there a chance there are any issues with the dependencies of the wrapped dataset?
hmm, does not seem to be the case
my trick with these sorts of issues is to open a notebook and try and do this config with the python api
Yes, that's what I meant by "I've (successfully) tested the config before" 😓
d
😞
debugger?
any chance the CLI is using the wrong env?
c
hmm, I get the same error If I use
pandas.CSVDataSet
as dataset
Copy code
optical_images_raw:
    type: PartitionedDataset
    path: data/01_raw/optical/
    dataset: pandas.CSVDataSet
    filename_suffix: .csv
    overwrite: True
n
DataSetError: Class ‘PartitionedDataset’ not found or one of its dependencies has not been installed.
Uhhh, at this point we should be able to surface the underlying error better. Can you try importing the underlying dataset directly from a terminal?
Any problem with
Copy code
import my_pkg.extras.datasets.image_dataset.ImageDataSet
c
yes, it works
it does not seem to come from the dataset, I get the same error with
pandas.CSVDataSet
n
wait, can you try
PartitionedDataSet
instead of
PartitionedDataset
?
👍 1
This rename happens in the last few months. I am checking which version did we change this now, probably should update the release notes
Okay, I think the oldest version start changing to
PartitionedDataset
is 0.18.11
PartitionedDataSet will still work for the entire 0.18.x series, in 0.19 we will have lower case consistently.
Patching the release note now.
c
ok, let me try
If you can, maybe just upgrade to 0.18.11+
c
yep
thanks! working!
d
😅
c
guys, that's wonderful to have a responsive and knowledgeable community! This kind of thing can drive you crazy otherwise. Big 🤗 to you
thankyou 4
K 4
❤️ 3
🎉 3
cc @Roberto P. Palomares (fyi)
🥳 1