데이터 로드

 

read_pandas_dataframe

Creates a new Dataflow based on the contents of a given pandas DataFrame.

​

Parameters

df

pandas DataFrame to be parsed and cached at 'temp_folder'.

​

temp_folder

path to folder that 'df' contents will be written to.

​

overwrite_ok

If temp_folder exists, whether to allow its contents to be replaced.

​

in_memory

Whether to read the DataFrame from memory instead of persisting to disk.

​

Returns

Dataflow that uses the contents of cache_path as its datasource.

​

Remarks

If 'in_memory' is False, the contents of 'df' will be written to 'temp_folder' as a DataPrep DataSet. This folder must be accessible both from the calling script and from any environment where the Dataflow is executed.

If the Dataflow is guaranteed to be executed in the same context as the source DataFrame, the 'in_memory' argument can be set to True. In this case, the DataFrame does not need to be written out. This mode will usually result in better performance.

​

참고

The column names in the passed DataFrame must be unicode strings (or bytes). It is possible

to end up with Integer types column names after transposing a DataFrame. These can be

converted to strings using the command:

df.columns = df.columns.astype(str)

import azureml.dataprep as dprep
// dprep.read_sql 써서 하려니까 잘 안됨.. 
// dataframe 에서 dataflow로 변경할 때 dprep.read_pandas_dataframe(df=df, in_memory=True)

 

 

https://docs.microsoft.com/ko-kr/python/api/azureml-dataprep/azureml.dataprep?view=azure-ml-py

azureml.dataprep package - Azure Machine Learning Python

https://docs.microsoft.com/ko-kr/azure/machine-learning/service/how-to-load-data

로드: Data Prep Python SDK - Azure Machine Learning service