Hi @385653
It works from single user clusters using dbfs path.
On Shared clusters, please set the spark conf at the notebook level where you would convert the json content into base64 string. This is a workaround as shared clusters do not support dbfs fuse(/dbfs). Example for Shared clusters below.
Ex :
spark.conf.set("credentials", base64_string)
df = spark.read.format("bigquery").option("credentialsFile", base64_string).option("table", "database.test").load()
display(df)
Volumes are the way going forward. However, this particular usecase to read crednetialfile is not supported. Engineering is working on it. There is no ETA at the moment. Once, Volumes are supported, you can use volume path for both Single User and Shared Clusters instead of dbfs going forward.
