BUG: How to read local csv file

[YishuiLi]

Describe the bug

Create a cluster with one supervisor and two workers on CentOS7, then connection this cluster in windows 10 and read local csv file with xorbits.pandas.read_csv, but got FileNotFoundError exception. How to read local csv file?

To Reproduce

1. Python 3.8.18

2. Xorbits 0.7.1

3. numpy and pandas a. pandas 2.0.3 b. numpy 1.24.4

4. Full stack of the error.

5. Minimized code to reproduce the error.

   
   import xorbits
   import xorbits.pandas as pd

xorbits.init('http://192.168.2.130:7005') df = pd.read_csv(r"F:\dataset\iris.csv", header=0, sep=",", encoding="utf-8") xorbits.run(df)

[qinxuye]

FileNotFoundError, can you check if the file exists?

[YishuiLi]

FileNotFoundError, can you check if the file exists?

File exists in the local computer, but not in the cluster. How to read this local csv file? Does every node require file transfers?

[qinxuye]

You need to store the file in the object store, like hdfs or oss etc, ensure the file can be seen in the cluster.

[YishuiLi]

You need to store the file in the object store, like hdfs or oss etc, ensure the file can be seen in the cluster.

Got it. Thanks!