The reporting_entity field shows that about 15% of rows have changed their entity code between the first and second half of the year, which is unusual. If the same location can report under different codes, that creates a matching problem in the time series. Should we standardize the entity codes, or should we note the instability?