Payment to yourself

[fabiorem]

I was transfering bitcoin between two wallets, and generating legacy addresses through the console.
However, without noticing it, I generated a address to the wallet which would be sending bitcoin, instead of the one receiving it. I just forgot to select the other wallet in bitcoin-qt.
So the bitcoin was sent to the same wallet, and now have 0 confirmations forever. It appears as "n/a" in the transaction ledger, instead of the address. Also a segwit change address was generated.
Minutes later I tried to send this bitcoin (0.12 actually) from these both adresses (the legacy and segwit) to another new legacy adress in the other wallet. This time I paid attention to it and selected the right wallet.
However, these are both unconfirmed as well. Blockchain.com also shows the first wrong transaction as unconfirmed.
Will these two transactions be confirmed, since the first one (payment to yourself) was not? Or did I lose the money due to this mistake?

[mocacinno]

It shouldn't be a problem to create a transaction funding an addres generated by the same wallet... The reason why your transaction remains unconfirmed must be something else: insufficient fees? Not broadcasted? invalid transaction?
EDIT: i just realised you looked at your transaction(s) on a block explorer, so they're probably broadcasted... Invalid Insufficient fees is still an option tough, or maybe the unspent output was used in 2 transactions? Hard to say without knowing the tx id.

I had a hard time following the rest of your story, but if you use the unspent output generated by an unconfirmed (parent) transaction to create a child transaction, the parent transaction HAS to be confirmed first (or at the same time) as the child transaction. It's impossible for the child transaction to be confirmed if the parent is unconfirmed. Both parent and child ending up in the same block is fine tough...

Last point: if i understand your story correctly, it looks like you didn't lose your funds... Backup your wallet.dat and try abandontransaction from the gui, or (if everything else fails), start your wallet with -zapwallettxes

OR... Alternatively, export your private keys and import them into electrum... Do realise that doing this might be a bad idear security-wise... It's just not a good idear to start manipulating unencrypted private key(s), so i'd personally consider your wallet to be potentially exposed and start a new wallet just to be sure...

EDIT: fixed a typo...

[Charles-Tim]

<...>

Never reveal your private key to anyone, but can you tell us the txids? If not, what fee did you use to make the transactions (both the first and the second transactions)? The fee will like be the issue as suggested by mocacinno. The fee used for CPFP can also be very smal leading to unconfirmed transactions.

[ranochigo]

However, without noticing it, I generated a address to the wallet which would be sending bitcoin, instead of the one receiving it. I just forgot to select the other wallet in bitcoin-qt.

That is correct. Transactions to an address that is in the same wallet shows up as N/A on the address column.

So the bitcoin was sent to the same wallet, and now have 0 confirmations forever. It appears as "n/a" in the transaction ledger, instead of the address. Also a segwit change address was generated.

Not forever, likely just unconfirmed.

Minutes later I tried to send this bitcoin (0.12 actually) from these both adresses (the legacy and segwit) to another new legacy adress in the other wallet. This time I paid attention to it and selected the right wallet.
However, these are both unconfirmed as well. Blockchain.info also shows the first wrong transaction as unconfirmed.
Will these two transactions be confirmed, since the first one (payment to yourself) was not? Or did I lose the money due to this mistake?

My guess is that what you've done is that you've spent the outputs of the first transaction in the second transaction. This means that your second transaction cannot be confirmed without the first transaction confirming. This is not an issue however. If you've signaled RBF, you can try to use abandon transaction as mentioned for both transaction and spend it again with a higher fee if it is not in your mempool and it likely is. 

I'm not sure if Bitcoin Core allows the RBF to have different outputs. IIRC, it doesn't so that could be a problem. Should try to create a raw transaction in that case, with care.

start your wallet with -zapwallettxes

Zapwallettxes was removed in 0.21.0, just in case OP tries it.

[mocacinno]

--snip--
Zapwallettxes was removed in 0.21.0, just in case OP tries it.

ouch... Didn't know that... I'm still running an older version.
To bad, this startup parameter was quite usefull imho, i've used it several times.

Thanks for the update

[BitMaxz]

Do you have any details about your transaction fee?

If the first transaction has a low transaction fee even you send them back to your wallet you still need to wait for this transaction to be confirmed first before your 2nd transaction confirmed.
And take note it also depends on the fee you set on both 1st and 2nd transactions.

Recently there is a big spike in the Mempool stats which is one of the reasons why your transaction still unconfirmed look at this below.



I suggest you double-spend the first transaction(Use same input and output) but with an additional fee, so that, your first transaction will be rejected.
If you already made a 2nd transaction it will also be rejected.

About the recommended fee use this https://jochen-hoenicke.de/queue/ scroll down and look for "Mempool size in MB" just point your mouse to the recent graph always use a fee which is under 1MB don't use the fee above the 1MB.
The current recommended fee right now is 120sat with 0.408MB in size.

[fabiorem]

<...>

Never reveal your private key to anyone, but can you tell us the txids? If not, what fee did you use to make the transactions (both the first and the second transactions)? The fee will like be the issue as suggested by mocacinno. The fee used for CPFP can also be very smal leading to unconfirmed transactions.

I just forgot to check the wallet when generating the address, so I thought it was in the destination wallet when I sent it.
This is the wrong transaction I made to the same wallet:

https://www.blockchain.com/pt/btc/tx/4158a8721afdf3c26e09aea846a5f5857cfca715dcba014954a3829e8bc2e38b

You see there is a segwit address generated for the change.
Since I thought it would not be confirmed (as it was inside the same wallet), I re-sent from these two address to another two new addresses:

https://www.blockchain.com/btc/tx/8d3e40de7570ec8593dcc09d06f9336c9b7425a8b025e4d8174e2a6ea4d80719
https://www.blockchain.com/btc/tx/39e1201781797e8d6c3fe95ddef24ff381841a06ab302b151328d5ab0fee12b0

I was doing this clean up because the original address (the one starting with 19WLod) had too many transactions, and I read quantum computers could attack P2PKH addresses with more than one transaction, so I started to slowly transfer between wallets two weeks ago. But this time, I had this oddity of "sending to yourself". I never saw this before. Correct me if I'm wrong, do sending to your own wallet need confirmations? Maybe I just made a mistake when I re-sent to these two new addresses, instead of waiting for it to confirm. I put 100 sat/byte for fee in each transaction, so maybe I'm still early to make any conclusion about it.

[mocacinno]

I'll try to give an answer to your questions... A bit unstructured, but still valid

It might sound extravagant, but those 100 sat/byte and 75 sat/byte are not sufficient at the moment. This is why your transaction remains unconfirmed (so my initial idear was correct)

Some of these transactions are indeed using an unspent output from an unconfirmed transaction (see my first post).

Yes, transactions withing your own wallet need to be confirmed to.

AFAIK, those quantum computer rumours are just FUD.

[fabiorem]

I suggest you double-spend the first transaction(Use same input and output) but with an additional fee, so that, your first transaction will be rejected.
If you already made a 2nd transaction it will also be rejected.

I cant, because, although I had used the RBF feature, I had spent the change address in the third transaction, so the RBF is no longer available in the first transaction.

[ranochigo]

I cant, because, although I had used the RBF feature, I had spent the change address in the third transaction, so the RBF is no longer available in the first transaction.

You can create a new raw transaction that spends the inputs from your first transaction and thus doing an RBF on the first and invalidate all of the subsequent transactions.

[Charles-Tim]

This is the wrong transaction I made to the same wallet:

https://www.blockchain.com/pt/btc/tx/4158a8721afdf3c26e09aea846a5f5857cfca715dcba014954a3829e8bc2e38b

With what I have checked, your transaction is just a normal transaction sent to two different addresses, no one to the same address. On the blockchain explorer:

You the sender use address 19WLodTpJkgJMZyQbpp3mDSLGdcaVCcQXf to make a transaction into two addresses 18s15o5fM1HvxhQEieBpcHKE5qzNLeSyHU and 3ETsCeNykiwUNFeojRA8isv55Lxx3xVkaS. You did not make any transaction into the same address. The transaction fee used is okay but for now low for transaction to be confirmed at present due to mempool congestion. Just wait for mempool to be less congested.

Since I thought it would not be confirmed (as it was inside the same wallet), I re-sent from these two address to another two new addresses:

https://www.blockchain.com/btc/tx/8d3e40de7570ec8593dcc09d06f9336c9b7425a8b025e4d8174e2a6ea4d80719
https://www.blockchain.com/btc/tx/39e1201781797e8d6c3fe95ddef24ff381841a06ab302b151328d5ab0fee12b0

The fee used for CPFP is low too is the reason why the transaction is not confirmed in time.

You can check it now, all the three transactions already have one confirmation. Your transactions are successful already.

[fabiorem]

These transactions received a confirmation each, right now.
I didn't lost the money, then.

Thanks for the information.

[Charles-Tim]

<...>

You made the mistake thinking you sent to the same bitcoin address, although it was different addresses, be it from your wallet or not. Supposing you did not make use of CPFP to spend the unconfirmed output from the first transaction, the transaction will still be confirmed at this time because the fee used for the second and third transaction were even lower than the fee used for the first transaction.

[fabiorem]

<...>

You made the mistake thinking you sent to the same bitcoin address, although it was different addresses, be it from your wallet or not. Supposing you did not make use of CPFP to spend the unconfirmed output from the first transaction, the transaction will still be confirmed at this time because the fee used for the second and third transaction were even lower than the fee used for the first transaction.

Not really. I knew it was different addresses. The issue was the "n/a" which appeared in the ledger, which made me think it would not need any confirmation for it. But in fact, it does need confirmation, as any other transaction.