It is rather difficult to calculate the reputation exactly from SEDE1, but the total score (i.e., the sum of the scores of all user's posts) can serve as a reasonable proxy.
Here is a query which looks at such users: Deleted accounts above the given score and number of posts. Notice that you can change the parameters minscore and minposts - the query returns users who have both score and number of posts $\ge$ the given values. For example, you can set on of them, or even both, to zero - to remove the restriction. (As you have already mentioned, once you have userid, you can find more about that user, such as the list of their posts: Is there a way to check all questions/answers posted from a deleted account. I will add that sometimes you might be able to find also a username that the user had before the deletion - the methods mentioned here work also for deleted users: Recognizing identifiable users changing user names.)
You can make a similar query which lists both existing and deleted users: Users above the given score and number of posts. Of course, if we're making a query combining both deleted and existing users, we can include also some data from the table Users for the ones which are not deleted, like in this query.
Re: I have no idea how to find the userid of someone I used to know but have disappeared. I will also add that if you have a post or a comment from a deleted user, you can see the userid there. You will find userxyz - where xyz is the userid - in places where you see username for the accounts which are not deleted.
If you remember the username and you want the userid, you might simply try to search where the username was mentioned on MO - which might lead you to some of their posts or comments. Or you can search for comments mentioning this username - and you might be lucky enough that you'll find a comment responding to a comment from the delete user. Examples of such queries are here and here.
1Some discussions related to calculating reputation from SEDE: How to retrieve exact reputation on "legendary day"?, Reputation History from Data Explorer, A query, days hitting rep cap vs days a member (rep).