Multi Mapping in Dapper. Receiving the error in SpiltOn - dapper

Question

*Can You explain the Split on function in the multimap *
I am Trying to get the data from the Database using Dapper ORM. I have received the following error
System.ArgumentException : When using the multi-mapping APIs ensure you set the splitOn param if you have keys other than Id
Parameter name: splitOn
public abstract class Domain
{
public Guid Id { get; set; }
}
public abstract class ItemBase : Domain
{
private IList<Image> images = new List<Image>();
public Guid? ParentId { get; set; }
public string Name { get; set; }
public IList<Image> Images { get { return images; } }
}
public class Meal : ItemBase
{
}
public class Item : ItemBase
{
private IList<Meal> meals = new List<Meal>();
public IList<Meal> Meals { get { return meals; } };
}
public class Image : Domain
{
public byte Img { get; set; }
public string Description { get; set; }
}
public class MealImageLink : Domain
{
public Guid ItemId { get; set; }
public Guid ImageId { get; set; }
}
/* search function to take dat from the table */
private List<Meal> SearchMeals(Guid id)
{
var query = #"SELECT meal.[Name],meal.[Description],meal.
[Price],mealImage.[Image] as Img
FROM [MealItems] as meal
LEFT JOIN [MealImageLink] mealImageLink
on meal.Id= mealImageLink.MealItemId
LEFT JOIN [Images] mealImage on
mealImageLink.ImageId=mealImage.Id
WHERE meal.[ParentId]=#Id";
List<Meal> meals = ( _connection.Query<Meal, MealImageLink, Image, Meal>
(query, (meal, mealLink, mealImage) =>
{
meal.Images.Add(mealImage);
return meal;
}, new { #Id = id })).ToList();
return meals;
}

Answer 1

The multi-map feature is really more intended for scenarios like:
select foo.Id, foo.whatever, ...,
bar.Id, bar.something, ...,
blap.Id, blap.yada, ...
from foo ...
inner join bar ...
left outer join blap ...
or the lazier but not uncommon:
select foo.*, bar.*, blap.*
from ...
inner join bar ...
left outer join blap ...
But in both of these cases, there is a clear and obvious way to split the horizontal range into partitions; basically, whenever you see a column called Id, it is the next block. The name Id is configurable for convenience, and can be a delimited list of columns for scenarios where each table has a different primary key name (so User might have UserId, etc).
Your scenario seems quite different to this. It looks like you're currently only selecting 4 columns with no particular way of splitting them apart. I would suggest that in this case, it is easier to populate your model via a different API - in particular, the dynamic API:
var meals = new List<Meal>();
foreach(var row in _connection.Query(sql, new { #Id = id }))
{
string name = row.Name, description = row.Description;
decimal price = row.Price;
// etc
Meal meal = // TODO: build a new Meal object from those pieces
meals.Add(meal);
}
The dynamic API is accessed simply by not specifying any <...>. With that done, columns are accessed by name, with their types implied by what they are being assigned to - hence things like:
decimal price = row.Price;
Note: if you want to consume the row data "inline", then just cast as soon as possible, i.e.
// bad: forces everything to use dynamic for too long
new Meal(row.Name, row.Description, row.Price);
// good: types are nailed down immediately
new Meal((string)row.Name, (string)row.Description, (decimal)row.Price);
Does that help?
Tl;dr: I just don't think multi-mapping is relevant to your query.
Edit: here's my best guess at what you intend to do - it simply isn't a good fit for multi-map:
var meals = new List<Meal>();
foreach (var row in _connection.Query(query, new { #Id = id })) {
meals.Add(new Meal {
Name = (string)row.Name,
Images = {
new Image {
Description = (string)row.Description,
Img = (byte)row.Img
}
}
});
}
return meals;