如何在Databricks中使用Pandas的pd.read_excel读取/FileStore/tables/目录下的Excel文件及解决FileNotFoundError问题

I’ve run into this exact issue before, so let’s break down the steps to get your Excel file loaded into a pandas DataFrame successfully.

Step 1: Double-check the exact filename in DBFS #

When you upload files via the Databricks UI, the platform often appends a random alphanumeric suffix to the filename (like abc_12345.xlsx) to avoid overwriting existing files. This is probably why your path isn’t working even though you see the file in the UI.

To confirm the actual filename, run this command in a notebook cell:

dbutils.fs.ls("/FileStore/tables/")

Look for your file in the output—note the full name including any suffixes. Then update your pandas read command with the correct path, e.g.:

import pandas as pd
df = pd.read_excel("/dbfs/FileStore/tables/abc_12345.xlsx")
display(df)

Step 2: If direct DBFS access still fails, copy to a local temp directory #

In some cluster configurations, pandas might have trouble reading directly from the DBFS mount. A reliable workaround is to copy the file to Databricks' local temporary storage first, then read it:

# Copy the file from DBFS to local tmp directory
dbutils.fs.cp("/FileStore/tables/abc_correct_name.xlsx", "file:/tmp/abc.xlsx")

# Read from local tmp
import pandas as pd
df = pd.read_excel("/tmp/abc.xlsx")
display(df)

Step 3: Read directly into memory (no local file copy) #

If you prefer not to create a local copy, you can read the file content into a BytesIO object and pass it to pandas:

from io import BytesIO
import pandas as pd

# Read the file content as bytes
file_content = dbutils.fs.cp("/FileStore/tables/abc_correct_name.xlsx", BytesIO())

# Reset the BytesIO pointer to the start
file_content.seek(0)

# Load into pandas DataFrame
df = pd.read_excel(file_content)
display(df)

All these methods are pure Python, so they should work within your organization's restrictions on Scala code.

内容的提问来源于stack exchange，提问作者Chaitanya Vardhan