如何正确输入dask Dataframe作为函数的参数？

Question

When i am passing a Dask Dataframe as parameter he is converting to pandas Dataframe.

print(type(sellout_df))
simulate_sku_predictions(sellout_df.loc[(sellout_df['sku'] == sku) & (sellout_df['store_id'] == store)].compute(), store, sku)

Prints => <class 'dask.dataframe.core.DataFrame'>

Entering on defined function

def simulate_sku_predictions(sellout_sku_df, store, sku):
    print(type(sellout_sku_df))

Prints => <class 'pandas.core.frame.DataFrame'>

I can't use compute and dask functions.

I'm new to dask but i don't think it's appropriate converting in the middle of the code if "i don't have to".

Answer 1

dask.DataFrame.compute() returns a Pandas DataFrame . Thus, the code is not passing a Dask DataFrame to simulate_sku_predictions . The argument,

sellout_df.loc[(sellout_df['sku'] == sku) & (sellout_df['store_id'] == store)].compute()

is evaluated to a Pandas DataFrame before it is passed as an argument to simulate_sku_predictions .

If you remove the call to compute() , then sellout_df.loc[...] will be a Dask DataFrame, and you could pass that to simulate_sku_predictions .