Loading data from pickle files
Hi,
I have some pandas dataframes that are saved to files outside Dataiku. I have put the files into a managed folder, and would like to import them as pandas dataframes.
I have tried different variations like this:
folder = dataiku.Folder("folder_name")
with folder.get_download_stream('df_filename.pkl') as f:
model = f.read()
This just leaves me with a binary variable, "model", that I can't convert to a dataframe. I have tried pickle.load(f) and pickle.load(f.read), but that does not help. Is there anyone that have experience with this?
Thanks,
Anders
Answers
-
Turribeach Dataiku DSS Core Designer, Neuron, Dataiku DSS Adv Designer, Registered, Neuron 2023 Posts: 2,067 Neuron
This post shows how to load a pickle file:
https://community.dataiku.com/t5/Using-Dataiku/Importing-a-pickle-model-in-Dataiku/m-p/33528
But I am not sure what you are trying to do. A pickle file is a binary model file. What exactly are you trying to achieve?
-
Thanks,
What I try to do is to load a saved dataframe and use it as a dataframe in Dataiku. In python I can do something like:
df = joblib.load(nameOfSavedFile)
This will give me a pandas dataframe (df) that I can use. But loading the file into Dataiku leaves me with the binary content of the file, which I haven't found out how to convert to a dataframe.