比较两个表的最快方法是什么？

Question

For example there are two tables with same schema, but different contents:

Table1

| field1  | field2      | field3       |
----------------------------------------
| 1       | aaaaa       | 100          |
| 2       | bbbbb       | 200          |
| 3       | ccccc       | 300          |
| 4       | ddddd       | 400          |

Table2

| field1  | field2      | field3       |
----------------------------------------
| 2       | xxxxx       | 200          |
| 3       | ccccc       | 999          |
| 4       | ddddd       | 400          |
| 5       | eeeee       | 500          |

---

The expected comparison result would be:

Deleted in B:

| 1       | aaaaa       | 100          |

Mismatch:

Table1:| 2       | bbbbb       | 200          |
Table2:| 2       | xxxxx       | 200          |
Table1:| 3       | ccccc       | 300          |
Table2:| 3       | ccccc       | 999          |

Newly added in B

| 5       | eeeee       | 500          |

Using C#， what is the fastest way to compare two tables?

Currently my implementation is: Check if each row in table1 has an exact match in table2; Check if each row in table2 has an exact match in table1.

The efficiency is n*n so for 100k rows it takes 20 mins to run on a server.

Many thanks

Answer 1

You can try something like this, should be quite fast:

class objType
{
    public int Field1 { get; set; }
    public string Field2 { get; set; }
    public int Field3 { get; set; }

    public bool AreEqual(object other)
    {
        var otherType = other as objType;
        if (otherType == null)
            return false;
        return Field1 == otherType.Field1 && Field2 == otherType.Field2 && Field3 == otherType.Field3;
    }
}

var tableOne = new objType[] {
    new objType { Field1 = 1, Field2 = "aaaa", Field3 = 100 },
    new objType { Field1 = 2, Field2 = "bbbb", Field3 = 200 },
    new objType { Field1 = 3, Field2 = "cccc", Field3 = 300 },
    new objType { Field1 = 4, Field2 = "dddd", Field3 = 400 }
};

var tableTwo = new objType[] {
    new objType { Field1 = 2, Field2 = "xxxx", Field3 = 200 },
    new objType { Field1 = 3, Field2 = "cccc", Field3 = 999 },
    new objType { Field1 = 4, Field2 = "dddd", Field3 = 400 },
    new objType { Field1 = 5, Field2 = "eeee", Field3 = 500 }
};

var originalIds = tableOne.ToDictionary(o => o.Field1, o => o);
var newIds = tableTwo.ToDictionary(o => o.Field1, o => o);

var deleted = new List<objType>();
var modified = new List<objType>();

foreach (var row in tableOne)
{
    if(!newIds.ContainsKey(row.Field1))
        deleted.Add(row);
    else 
    {
        var otherRow = newIds[row.Field1];
        if (!otherRow.AreEqual(row))
        {
            modified.Add(row);
            modified.Add(otherRow);
        }
    }
}

var added = tableTwo.Where(t => !originalIds.ContainsKey(t.Field1)).ToList();

Might be worth overriding Equals instead of AreEqual (or making AreEqual a helper method outside the class definition), but that depends on how your project is setup.