Zcash dataset ignores Sapling pool when calculating fees about bitcoin-etl

[Mitchellpkt]

@medvedev1088 Thanks! Looks great, seems to be returning correct values now. :- )

[Mitchellpkt]

As a quick note from my own EDA:

• Transparent >> Transparent fee is calculated correctly
• Shielded >> Shielded always shows fee=0
• Transparent >> Shielded produces anomalous value (easy workaround: SELECT ... ABS(ABS(input_value) - ABS(output_value)) as real_fee ...)
• Transparent >> {mixed} produces anomalous value that is not as straightforward to correct

[medvedev1088]

@str4d It seems I missed the negation here:

if transaction.value_balance > 0: 
    input = BtcTransactionInput() 
    input.type = ADDRESS_TYPE_SHIELDED 
    input.value = transaction.value_balance 
    transaction.add_input(input) 
if transaction.value_balance < 0: 
    output = BtcTransactionOutput() 
    output.type = ADDRESS_TYPE_SHIELDED 
    output.value = -transaction.value_balance              <---------- the minus sign is missing
    transaction.add_output(output) 

With this fix the fee calculation should be correct. Please confirm.

@Mitchellpkt could you provide an example for this case

Transparent >> {mixed} produces anomalous value that is not as straightforward to correct

[medvedev1088]

Fixed in BigQuery dataset. Let me know if you find any other issues.