@Gustavo Martins​ :

I see, thank you for the clarification. In that case, you can use the spark.conf.get("spark.sql.warehouse.dir") method to get the location of the default database directory, which is typically the root directory of the Databricks File System (DBFS). From there, you can construct the path to the table's metadata file, which includes the schema information.

Here's an example:

target_table_name = 'whatevertable'
 
@dlt.table(
     name=target_table_name
)
def update_target():
    dbfs_root = spark.conf.get("spark.sql.warehouse.dir")
    target_table_path = f"{dbfs_root}/{target_table_name}"
    target_table_metadata = spark.read.format("parquet").load(f"{target_table_path}/_metadata")
    target_table_schema = target_table_metadata.schema.simpleString()
 
    max_date = (
            spark
            .table(f"`{target_table_schema}`.`{target_table_name}`")
            .select(F.max('reference_date'))
    )
 
    ...

This code reads the _metadata file of the target table to get its schema information, which is then used to construct the full target table path. You can then use the full schema-qualified table name to access the table with the

spark.table() method.

Note that the spark.sql.warehouse.dir configuration property may not be set in some environments, in which case you may need to manually specify the root directory of your default database.