PostgreSql and EF – Getting ‘duplicate key’ error while doing ‘ToDictionary’ after grouping

Here is my query. UserName column is citext, having non-unique index on it.

var logs = Db.Logs
            .GroupBy(x => x.UserName)
            .ToDictionary(g => g.Key, g => g.Max(m => m.Time));

I’m getting an exception:

System.ArgumentException: An item with the same key has already been added

I don’t understand, how is it possible to be any dupe keys in resultant dictionary, if grouped values can’t have dupes by design?

---

P.S. I have an update on this. While investigating, I’ve looked into generated sql for GroupBy statement and got stuck (see below). Why group by interpreted as order by instead??

SELECT x."Time", x."UserName"
FROM "UserAccessLogs" AS x
ORDER BY x."UserName" NULLS FIRST

Answer

Well, seems that is a kind of ‘documented behavior‘. Grouping for further manipulation (like ToDictionary f.e.) requires anonymous type representation before. I rewrote my query as

var logs = Db.Logs
            .GroupBy(x => x.UserName)
            .Select(g => new { UserName = g.Key, Time = g.Max(m => m.Time)})
            .ToDictionary(d => d.UserName, d => d.Time);

and it works as expected. That’s why EF was interpreting my query wrong and that’s why I was getting a ‘dupe keys‘ error.