You are using Azure Synapse Analytics serverless SQL pool to query Parquet files in Azure Data Lake Storage Gen2. You notice that queries are slower than expected. The files are large (500 MB each) and not partitioned. You need to improve query performance without moving data. What should you do?