Serilog: difference between {..} and {@..}

.net, serilog

Solution

`{d1}` converts unrecognized types like the anonymous one here into `string`s for logging, i.e. using `ToString()`. So your log event in the first example will end up with a property like (here in JSON):

{
  "d1": "{ x = 5, y = 88 }"
}

Using `{@d1}` will cause the parameter to be serialized as structured data:

{
  "d1":
  {
    "x": 5,
    "y": 88
  }
}

Where appropriate, the second example is much more useful for manipulation/analysis.

The reason for this "opt-in" requirement is that most types in .NET programs convert nicely into strings, but aren't cleanly/meaningfully serializable. By opting in to serialization with `@` you're saying: "I know what I'm doing, serialize this object!" :)

Problem

Given this code: ``` var d1 = new { x = 5, y = 88 }; Log.Logger.Information("{d1}", d1); Log.Logger.Information("{@d1}", d1); ``` How will the object in d1 be logged differently in the two Log.Logger.Information(...) lines? In other words, what is the effect of adding the @ in between the { } ? I read https://github.com/serilog/serilog/wiki/Structured-Data under the heading "Preserving Object Structure", but that didn't make sense to me.

Original source