How to Import Data from Hive Table into TIBCO ComputeDB Table
Option 1
If Hive tables have data stored in Apache Parquet format or Optimized Row Columnar (ORC) format the data can be copied directly into TIBCO ComputeDB tables.
For example,
CREATE EXTERNAL TABLE <hive_external_table_name> USING parquet OPTIONS(path path-to-parquet-or-orc)
CREATE TABLE <table_name> USING COLUMN AS (select * from hive_external_table_name)
For more information on creating an external table, refer to CREATE EXTERNAL TABLE.
Option 2
Take the RDD[Row] from Dataset of Hive Table and insert it into column table.
For example,
val ds = spark.table("Hive_Table_Name")
val df = snappy.createDataFrame(ds.rdd, ds.schema)
df.write.format("column").saveAsTable("Snappy_Table_Name")
In above example, 'spark' is of type SparkSession and 'snappy' is of type SnappySession.