Support for TTL per insertion while using DataFrame.save method

We want to add support of using "ttl" filed . For example currently spark supports TTL field using

rdd.saveToCassandra("test", "tab", writeConf = WriteConf(ttl = TTLOption.perRow("ttl")))

where rdd is something like ( last field is TTL in seconds)
val rdd = sc.makeRDD(Seq(
KeyValueWithTTL(1, 1L, "value1", 100),
KeyValueWithTTL(2, 2L, "value2", 200),
KeyValueWithTTL(3, 3L, "value3", 300)))

We want do similar with dataframe apis like
data.write
.format("org.apache.spark.sql.cassandra")
.options(Map( "table" -> tableName, "keyspace" -> dbName,
"spark_cassandra_connection_host" -> createStringList(hosts),
"spark_cassandra_output_metrics" -> "false",
"spark_cassandra_output_concurrent_writes" -> "20",
))
.mode(SaveMode.Append)
.save()
We want to use "ttl" field in our dataframe and want cassandra to apply it per row.