ComputeDB Data Source Limitations

This section lists the general limitations related to ComputeDB.

•

In order to introspect an empty schema, you must check the option “Detect New Resources During Re-Introspection” in the Data Source Introspection dialogue box.

•

Numeric functions pi() and cot() are not supported.

•

Queries resulting in a Null scan will not be pushed down.

•

All table identifiers must be specified in uppercase.

•

For statistics to work as expected, modify the following files while setting up ComputeDB:

- In the spark-defaults.conf file, add the line "spark.sql.thriftServer.incrementalCollect true"

- In the leads.conf file, add the line "-spark.driver.maxResultSize=0" - "0" (zero) is unlimited.

Note: The above 2 files can be found in the folder $SNAPPY_HOME/conf

•

If the Transaction Isolation property of the datasource is set as Read Committed, the following functions are not recognized:

- ascii()

- upper()

- concat()

- chr()