Unknown Address Format Parsing Bitcoin Addresses from Blockchair Data

Question

I've obtained a list of all current Bitcoin addresses with balances from Blockchair, which is downloadable from their data repository located here. I have downloaded the Bitcoin / Addresses TSV database dump.

The list consists of:

• ~23m P2PKH (1xxx) addresses
• ~5.8m P2SH (3xxx) addresses
• ~1.3m Bech32 (bc1xxx) addresses

However, there are also ~400k addresses that have prefixes that I don't recognise, including:

• d-xxx (example: d-b697921d996831d1a26ca8d8e92571c1)
• m-xxx (example: m-942943e227462513f33c4f32ec6f1e56)
• s-xxx (example: s-2ad548f2459c0ed6123400e5d61a479e)

Are these valid Bitcoin addresses?

If they are, how do I view information about them?

If they are not, any idea why they appear in the dump of current Bitcoin addresses from Blockchair?

Answer 1

Having investigated this following @MCCCS answer, the following is information about the addresses provided by Blockhair support:

In fact, there are no addresses on the Bitcoin blockchain. There are only scripts, some of them are standard and can be converted to a Bitcoin address, some of them can’t — like nulldata (OP_RETURN) outputs, nonstandard, and bare multisig. We use the d- prefix for nulldata, m- for bare multisig, and s- for everything else.

This is briefly mentioned in this API documentation section: https://blockchair.com/api/docs#link_300

Therefore, these addresses (for want of a better word) are valid transactions on the Bitcoin blockchain.

The prefixes used are:

• d- nulldata
• m- bare multisig
• s- everything else

In conclusion, the data is valid but the outputs are not addresses in the normal sense of the word.

A link to the issue I raised on the Blockchair GitHub is here.

Answer 2

Those are invalid addresses that don't belong to any cryptocurrency. You should file a bug report at https://github.com/Blockchair/Blockchair.Support/issues since those might even be their database secrets.

Those are of 16 bytes. My initial guess about their format would be SHA-1, but Bitcoin does not use SHA-1. Probably it's related to their database.

For example on this page https://blockchair.com/dogecoin/address/942943e227462513f33c4f32ec6f1e56 it seems Blockchair thinks 942943e227462513f33c4f32ec6f1e56 is a valid Doge address. It's not.

To prevent this from happening again in the future, they could implement input sanitization before INSERT statements