Hive Schema on Read vs Schema on Write

In traditional RDBMS a table schema is checked when we load the data. If the data loaded and the schema does not match, then it is rejected. This is called as Schema on write which means data is checked with schema when it written into the database. Let us take an example and look into this.

When we load the data our schema is checked, suppose we have 10 columns but data is loaded using 9 columns then schema is rejected. If first column is of INT type but first column of data is String type, then schema is rejected. This is called as schema on write, which means when we are writing the data at that time schema is enforced.

Hive supports Schema on read, which means data is checked with the schema when any query is issued on it. This is similar to the HDFS Write operation, where data is written distributedly on HDFS because we cannot check huge amount of data.

We cannot check each and every record of it as it will take months to check each and every record. This operation is fast and also improves performance.