[CALCITE-7136] Add Support for Druid Double Types

[s-robertshaw]

Using the std golang sql package, Scan functionality will quietly fail and the zero value will be used.

During the prepare phase of the sql request the correct types are returned.
xxx_hidden_ColumnClassName ="java.lang.Double"
xxx_hidden_Name ="DOUBLE"
xxx_hidden_Rep =Rep_NUMBER (22)

Special handling already exists for things such as floats and reals. Adding a doubles case solves the issue.

[mihaibudiu]

Is there a test you can write to show that this works?

[F21]

The driver supports Phoenix, HSQLDB and "Generic", with the code to map the columns for each type of database here:

• https://github.com/apache/calcite-avatica-go/blob/main/phoenix/phoenix.go
• https://github.com/apache/calcite-avatica-go/blob/main/hsqldb/hsqldb.go
• https://github.com/apache/calcite-avatica-go/blob/main/generic/generic.go

We do have tests against HSQLDB and Phoenix:

• https://github.com/apache/calcite-avatica-go/blob/main/driver_hsqldb_test.go
• https://github.com/apache/calcite-avatica-go/blob/main/driver_phoenix_test.go

However, we don't have tests for the generic mapping as there is no "generic" database to test against. If we test against a specific database, then we should have a mapping for that database, rather than use generic.

I am happy to accept the PR as it currently stands, but the commit message needs to be updated to the format: [CALCITE-XXXX] {JIRA issue title}