Robert Francis

```ex [ adapter_meta: %{ cache: #Reference, opts: [timeout: 15000, pool_size: 10, pool: DBConnection.ConnectionPool], pid: #PID, sql: AlogAdapter.Connection, telemetry: {UsingAlogAdapter.Repo, :debug, [], [:using_alog_adapter, :repo, :query]} }, opts: [], params: [], prepared:...

At first glance the arguments passed to our adapter look almost identical, with the **OBVIOUS** exception of the `:prepared` atom. > `:prepared` is the same as `:query` I just logged...

Hopefully this means that if we can somehow update the query that is passed to the adapter, to add subquery to it. Will look into possible approaches for this.

The code that will enable access to the all query can be found [here](https://github.com/RobStallion/alog_adapter/pull/3/files). At the moment I am unsure if this function is only called when `Repo.all` is called...

It appears that the reason for the error is because the `user changeset` function does not return an error when there is a duplicate email hash. That step looks like...

As there cannot be a unique index in alog (as many rows can relate to one user for example) it looks like alog is doing a version of this logic...

I need to think about this logic in a little more detail but it maybe be good to have check and doubleCheck saved in the state somewhere. If a king...

@SimonLab 👍 This pr resolved the issue for me as well

I have been looking into approaches that we could take to 'convert' `alog` into an `ecto adapter`. After looking at the [postgres implementation](https://github.com/elixir-ecto/ecto/blob/v2.2.8/lib/ecto/adapters/postgres/connection.ex), I think we will be able to...

Are we planning on turning the entire map/struct into a cid? If so how do we plan on comparing the new cid to the old one? Each CID should always...