What is the fastest way to import data from HDFS to Clickhouse? Spark is imported through JDBC or HDFS

hdfs Import the data on to clickhouse What is the fastest way ？spark adopt jdbc Import , still hdfs The table engine creates a ck surface , And then turn to mergeTree ？

---

Take the answer 1：

hdfs If the deposit is hive Table data ,ck Now the new edition can be used directly hive Table engine derivatives . If ck If the version doesn't work, use spark Derivative will be better . Others depend on the situation .（ This answer is collated from EMR OLAP(ClickHouse&StarRocks) User technical exchange group ）

---