Notice user that indexes are not applied because indexed data protocols are different.

[paryoja]

Describe the issue

I made an index using data path from "abfss://..../my_data". And executed a query using spark.sql(). Index was not applied to the query even though it is applicable.

I think it is because I created with the query below where location starts with "abfs://"

var query = "CREATE TABLE " + qualifiedTableName + s" $schema USING " + s"${format.toUpperCase()} " if (partitionColumns.nonEmpty) { query += s"PARTITIONED BY $partitionColumnsStr " } query += s"LOCATION '$location' "

Even though the paths points to the same data, due to the protocol difference, it is considered as different one.

Although it is safe consider abfs://.../my_data and abfss://.../my_data are different, is it possible to notify the users that the index is not applicated due to the differences? I think it is subtle so that some users can be confused due to this difference.

To Reproduce

Create table with abfs and build an index with abfss.

Expected behavior

Notice there are indexes for the same path but different protocol.

Environment

IDE : Azure synapse notebook Spark : 2.4 Platform : Azure synapse