dapper的经典方法是在查询枚举记录时使用字典存储主对象。
public IEnumerable<Place> SelectPlaces()
{
string query = @"SELECT p.id, p.PlaceName, t.id, t.tagname
FROM Place p INNER JOIN TagPlace tp ON tp.PlaceId = p.Id
INNER JOIN Tag t ON tp.TagId = t.Id";
var result = default(IEnumerable<Place>);
Dictionary<int, Place> lookup = new Dictionary<int, Place>();
using (IDbConnection connection = GetOpenedConnection())
{
// Each record is passed to the delegate where p is an instance of
// Place and t is an instance of Tag, delegate should return the Place instance.
result = connection.Query<Place, Tag, Place(query, (p, t) =>
{
// Check if we have already stored the Place in the dictionary
if (!lookup.TryGetValue(p.Id, out Place placeFound))
{
// The dictionary doesnt have that Place
// Add it to the dictionary and
// set the variable where we will add the Tag
lookup.Add(p.Id, p);
placeFound = p;
// Probably it is better to initialize the IEnumerable
// directly in the class
placeFound.Tags = new List<Tag>();
}
// Add the tag to the current Place.
placeFound.Tags.Add(t);
return placeFound;
}, splitOn: "id");
// SplitOn is where we tell Dapper how to split the record returned
// in the two instances required, but here SplitOn
// is not really needed because "Id" is the default.
}
return result;
}
