spark-avro ignores the compression option in DataFrameWriter

[juarezr]

Documentation says that should set compression in SparkConf:

spark.conf.set("spark.sql.avro.compression.codec", "deflate")

But others formats like parquet allows setting it in DataFrameWriter options:

DataFrameWriter<Row> writer = rowDataset.write() .format("com.databricks.spark.avro") .option("compression","snappy") .save(path);

For consistency, spark-avro could also respect this setting.