Spark Excel Library Insufficient Privileges

[jamesdavids]

Hi

We have a shared access mode cluster in which we have installed a maven library for reading excel files into a Spark DataFrame. When using an account with admin rights everything works fine, however when we run it as a standar user we always get `

org.apache.spark.SparkSecurityException: [INSUFFICIENT_PERMISSIONS] Insufficient privileges:`

We have managed to fix this by granting standard users the ANY FILE securable. However we would like to use a different approach since some installed libraries store data of all users in one common temp directory. These libraries might compromise user isolation. Do you know how can we make this work without granting ANY FILE to standard users? Thanks in advance.

[Walter_C]

Hello James, this issue seems to be related to shared cluster limitations, as per docs: Libraries used as JDBC drivers or custom Spark data sources on Unity Catalog-enabled shared compute require ANY FILE permissions.

Moving to a single user cluster must resolve the requirement of ANY FILE permission