File in Google Drive mounted as Directory in Collab

user_1897839 0
Viewed 1201

I am trying to access a CSV file hosted in my Google Drive in a Google Colab notebook. As described in the documentation, I mount my personal drive through the Web UI. The Drive content appears in the sidebar (almost) as expected. Not entirely, however, because the CSV file is shown as a directory in Google Colab.

This screenshot shows the CSV file in my Google Drive: Google Drive

I can download the file from Google Drive, and the content is the CSV file as expected.

This screenshot shows how the file appears in the Google Colab file browser like a directory, including a sub-directory 0.0.0: Google Colab tree

When running ls on the file inside Google Colab, the CSV file is also listed as directory, including the 0.0.0 sub-directory: Google Colab ls

Consequently, when I try to read the file in Python, it throws an IsADirectoryError: enter image description here

Another thing is that there are two more CSV files in the same directory in Google Drive (as seen in the screenshot above). They do not appear at all in the mounted drive in Google Colab.

There are a few questions about IsADirectoryErrors on SO, but the ones I have found are due to actual directories. As of writing, Google points me to [this questions]7, which has apparently been deleted.

Why is my file shown as a normal file in Google Drive, but as a directory in Google Colab? Where does the sub-directory 0.0.0 come from? Why are the two other files not visible in Google Colab?

google-colaboratory google-drive-api python
1 Answers

Try it this way using os library.

from google.colab import drive 
import os
import pandas as pd

Mount drive

drive.mount('/content/drive')

Print file names and note the file index

path = "/content/drive/MyDrive/data/educational_classifier"
fnames = os.listdir(path)
print(fnames)

Output should be: ['file1.csv', 'file2.csv', 'file3.csv']

Use the index to read file

df = pd.read_csv(os.path.join(path, fnames[index]))

e.g : df = pd.read_csv(os.path.join(path, fnames[3]))

user_52608720
Related