pip/conda install pyarrow 然後: df.to_parquet('somefile', engine='pyarrow') FastParquet在https://github.com/dask/fastparquet/的文件中說“只 ...
確定! 回上一頁