Database Reference
In-Depth Information
The preceding insert statement starts the MapReduce job and copies the data from
the Hive table to the HBase-backed table.
In the previous insert command, the number of records in both
Hive and HBase tables might not be the same, as HBase cannot have
duplicate row keys. Hence, the column that is mapped as rowkey to
the HBase table always provides the distinct values to HBase.
We can also map the HBase table to a single or multiple (used for distinct column
families) Hive external tables (by using external keyword with create statement).
In case of table-drop in Hive, external tables are not deleted, rather it removes
the metadata information about the table.
Summary
In this chapter, we learned about the different HBase clients that work synchronously
such as the REST client, HBase shell, Thrift client, Kundera—object mapper and
asynchronously such as Hive.
In the next chapter, we will look into HBase administration and cover HBase
monitoring and performance tuning as well.
