Databricks loader: Add collector_tstamp_date column

[istreeter]

When loading to delta tables, it is helpful to take advantage of data partitions for better query performance.

For snowplow data, an appropriate partitioning column is date of the collector timestamp. It is relatively low cardinality, and collector_tstamp is a field that appears in many analytic queries, including Snowplow's standard data models. To make this column available for partitioning, we need to explicitly add it to the COPY INTO statement.

[istreeter]

Closing in favour of #951