Daniel Thom

Cannot filter model years in query that uses comstock/resstock

3

copying from emails... **DAN:** Problem: Filtering model_year doesn't work with comstock/resstock because those datasets only have one year. If you query for model_year 2020, you will get nothing because 2018...

Time zone handling in queries causes confusion

4

Consider this example: Query asks for comstock data to be transformed to StandardScenarios. - StandardScenarios uses EST (no DST) and the time dimension query name is `time_est`. - dsgrid produces...

The code would need to exclude null values in each row.

When Spark writes CSV files, it creates a directory and makes one CSV per partition. Readers using Spark won't even notice. Pandas users might find it problematic.

Consider whether queries need to have an option to convert data to a specific time zone

3

Consider problems discussed in #192

Currently, we only support pivoted format.

The code in `TableFormatHandlerBase.convert_columns_to_query_names` will assert if the project time dimension has more than one column. If that rule is OK, we need to check this as registration time.

The code in PivotedTableHandler.add_columns will raise an exception if you do this. This probably isn't critical. If you add a column, such `state` when the base dimension is `county`, you...