DSS sets TIMESTAMP_TZ and TIMESTAMP_NTZ columns as strings in output datasets schemas. However, the DATE type is considered a date in an output dataset even though it doesn't have a time. TIMESTAMP_TZ columns do have both a time and a time zone but are set to a string type. The behavior seems to differ by recipe type (in version 11.1.2). SQL Script recipes result in DATE and TIMESTAMP_LTZ columns being set as dates and usable as dates elsewhere in DSS (e.g., in charts). SQL Query recipes seem to set DATE and TIMESTAMP_LTZ as dates but they show with a meaning of Date Unparsed and can't be used as dates in Charts.
It would be great if DSS consistently worked as follows across all recipe types:
1. TIMESTAMP_TZ and TIMESTAMP_LTZ (point in time types) were both set as dates in output dataset schemas and were directly usable as dates elsewhere in DSS (e.g., in Charts).
2. The user had the option to specify that DATE and TIMESTAMP_NTZ columns (types with no time and/or time zone) be treated as dates in output dataset schemas and be directly usable as dates elsewhere in DSS.
Additionally, it would good if the output dataset schema types that were set by DSS didn't generate an error message when reviewing the output dataset schema. For example, DSS sets a Snowflake DATE column to date in the schema but then reports the following error when one clicks on the Dataset Schema tab: "Type mismatch for column 2 (AS_OF_DT) : 'TIMESTAMP' in dataset, 'DATE'(91:DATE) in table".