-
Notifications
You must be signed in to change notification settings - Fork 9
Description
When exporting the RS_serbia dataset from ETLocal to ETSource, the parent shares of the following edges are determined through sparse graph queries:
industry_final_demand_for_metal_network_gas-industry_final_demand_for_metal_aluminium_network_gas@network_gas
industry_final_demand_for_metal_network_gas-industry_final_demand_for_metal_steel_network_gas@network_gas
The determined values are however not in balance:
industry_final_demand_for_metal_network_gas-industry_final_demand_for_metal_aluminium_network_gas@network_gas:
parent_share: 0.11863428388165434
industry_final_demand_for_metal_network_gas-industry_final_demand_for_metal_steel_network_gas@network_gas:
parent_share: 0.8813657161183456
Summed value = 0.99999999999999994
The first calculated parent share is rounded to 17 decimals, the second to 16 decimals. It's not clear to me why the second one is rounded to one less decimal. For now, we've manually corrected the second one to 17 decimals (changed to 0.88136571611834566), so that the shares add up to 1.0. Otherwise, the dataset won't load.
But every time the dataset is exported to ETSource, we need to be aware that this value is not changed back. We cannot do a quick fix in the dataset values since these shares are calculated based on EB values.
Next steps
I see some questions / follow up steps:
- Why does the second parent share round to less decimals? This seems to cause the imbalance
- Improve rounding of decimals method, less decimals would be preferred and in a way that the imbalance would not occur
- Alternatively, could we change the sparse graph queries to ensure that the values are balanced?