TIME (time64[ns]) type unhandled #505

QueryContext + rows.Next fails for a TIME column. The server returns the value as an Arrow time64 (nanoseconds, including when the SQL type is TIME with microsecond precision). The Arrow row scanner has no arm for time32 or time64, so the whole result fails with:

databricks: arrow row scanner unhandled type time64[ns]

The value is present in the Arrow batch. The failure is the scanner's type switch.

Reproducer, against a SQL warehouse:

SELECT TIME '09:15:00.123456'

rows, err := db.QueryContext(ctx, "SELECT TIME '09:15:00.123456'")
// err == nil
var dest any
ok := rows.Next()
err = rows.Scan(&dest)
// databricks: arrow row scanner unhandled type time64[ns]

Expected: a TIME value scans as a time-of-day (wall clock, no date and no location), and one unsupported column does not have to fail the result before any row is returned. DATE and TIMESTAMP already have arms in that switch https://docs.databricks.com/aws/en/sql/language-manual/data-types/time-type.