SqlBulkCopy from a List<>
Asked Answered
A

5

36

How can I make a big insertion with SqlBulkCopy from a List<> of simple object ?

Do I implement my custom IDataReader ?

Always answered 12/10, 2010 at 9:51 Comment(2)
Late to the party, but if you add this EntityDataReader class, there's an AsDataReader() extension method that does exactly that: github.com/matthewschrager/Repository/blob/master/…Intendance
(see new answer below for full implementation)Intendance
C
22

Simply create a DataTable from your list of objects and call SqlBulkCopy.WriteToServer, passing the data table.

You might find the following useful:

For maximum performance with SqlBulkCopy, you should set an appropriate BatchSize. 10,000 seems to work well - but profile for your data.

You might also observe better results when using SqlBulkCopyOptions.TableLock.

An interesting and informative analysis of SqlBulkCopy performance can be found here.

Clientele answered 12/10, 2010 at 10:5 Comment(0)
B
63

With FastMember, you can do this without ever needing to go via DataTable (which, in my tests, more-than-doubles the performance):

using(var bcp = new SqlBulkCopy(connection))
using(var reader = ObjectReader.Create(data, "Id", "Name", "Description"))
{
    bcp.DestinationTableName = "SomeTable";
    bcp.WriteToServer(reader);
}

Note that ObjectReader can also work with non-generic sources, and it is not necessary to specify the member-names in advance (although you probably want to use the ColumnMappings aspect of SqlBulkCopy if you don't specify them in the ObjectReader itself).

Belshazzar answered 7/6, 2013 at 9:53 Comment(6)
Excellent library! I just tried it right now and it works great.Blessed
I know this was from a few months back, but I am having a similar issue. It takes too long to load DataTable first, so I wanted to use this method. However is the strings listed in the params variable the actual names of the variables being used in order from the object the underlying data structure is iterating over?Avery
Scratch that-- figured it out and the answer was yes those are the names of the properties within the object.Avery
@Marc Gravell, when I use this with bcp and List<T>, I get error "Cannot convert source string to datetime". I'm guessing I need to add ColumnMappings, but to get column names I would have to do Reflection. Is there a way to get Property Names from the ObjectReader?Rhubarb
@Marc Gravell, when I grow up I want to code like you :). Thanks man you always provide great answersHanoverian
I tried your ObjectReader.Create(entities) and it was 8 times slower ??? then my new ObjectDataReader<T>(entities) using a Lazy<PropertyAccessor> from this site: codeproject.com/Articles/1095790/… HOW comes?Courses
C
22

Simply create a DataTable from your list of objects and call SqlBulkCopy.WriteToServer, passing the data table.

You might find the following useful:

For maximum performance with SqlBulkCopy, you should set an appropriate BatchSize. 10,000 seems to work well - but profile for your data.

You might also observe better results when using SqlBulkCopyOptions.TableLock.

An interesting and informative analysis of SqlBulkCopy performance can be found here.

Clientele answered 12/10, 2010 at 10:5 Comment(0)
I
12

Late to the party, but if you add this EntityDataReader class from Microsoft, there's an AsDataReader() extension method that does exactly that: https://github.com/matthewschrager/Repository/blob/master/Repository.EntityFramework/EntityDataReader.cs

(example [List].AsDataReader() implementation:)

var connStr = "";
using (var connection = new SqlConnection(connStr)) 
{
    var startTime = DateTime.Now;
    connection.Open();
    var transaction = connection.BeginTransaction();
    try
    {
        //var connStr = connection.ConnectionString;
        using (var sbCopy = new SqlBulkCopy(connection, SqlBulkCopyOptions.Default, transaction))
        {
            sbCopy.BulkCopyTimeout = 0;
            sbCopy.BatchSize = 10000;
            sbCopy.DestinationTableName = "Foobars";
            var reader = Foobars.AsDataReader();
            sbCopy.WriteToServer(reader);
        }
        transaction.Commit();
    }
    catch (Exception ex)
    {
        Console.WriteLine(ex.Message);
        transaction.Rollback();
    }
    finally
    {
        transaction.Dispose();
        connection.Close();
        var endTime = DateTime.Now;
        Console.WriteLine("Upload time elapsed: {0} seconds", (endTime - startTime).TotalSeconds);
    }
}
Intendance answered 23/4, 2016 at 22:30 Comment(1)
Some may need to add a column mapping for each column they want to import: sbCopy.ColumnMappings.Add(new SqlBulkCopyColumnMapping("col1", "col1"));Subdominant
C
1

Depending on what you are trying to accomplish by calling SqlBulkCopy in the first place, it might make more sense to use a Table-Valued Parameter (TVP). Using a TVP would make it trivial to send in a collection of any custom type. The data can be streamed in so you can avoid the DataTable (much like in @Marc Gravell's answer) and you can avoid SqlBulkCopy as well. TVP's allow for completely flexibility of how to handle the data once it gets to SQL Server as you call a Stored Procedure to pass the TVP data into and it appears as a Table Variable that you can do anything with, not just INSERT (which is the case with SqlBulkCopy). You can also get data back via a SqlDataReader, data such as newly created IDENTITY values. I added an example and some additional notes on this answer: How can I insert 10 million records in the shortest time possible?. And several years ago I wrote an article on SQL Server Central (free registration required), Streaming Data Into SQL Server 2008 From an Application, which is also noted in that linked answer, providing a working example of passing in a Generic List of a custom type, streamed in from a 3 million row text file.

Coffee answered 11/9, 2014 at 16:22 Comment(0)
R
0

Came across a similar situation trying to insert a couple million rows into the db.

Got this done by converting the List into a DataTable and then inserting the table into the database.

private static DataTable CreateDataTableItem(List<ListItem> ItemList)
    {
        DataTable dt = new DataTable();
        try
        {
            dt.TableName = "PartPrice";

            foreach (PropertyInfo property in typeof(ListItem).GetProperties())
            {
                dt.Columns.Add(new DataColumn() { ColumnName = property.Name, DataType = Nullable.GetUnderlyingType(property.PropertyType) ?? property.PropertyType, AllowDBNull = true });
            }

            foreach (var item in ItemList)
            {
                DataRow newRow = dt.NewRow();
                foreach (PropertyInfo property in typeof(ListItem).GetProperties())
                {
                    newRow[property.Name] = item.GetType().GetProperty(property.Name)?.GetValue(item, null) ?? DBNull.Value;
                }
                dt.Rows.Add(newRow);
            }
            return dt;
        }
        catch (Exception ex)
        {
            Console.WriteLine(ex.ToString());
            return null;
        }
    }

public class ListItem
{
    public int Id { get; set; }
    public string Name { get; set; }
    public int? NullableId { get; set; }

}

And then bulk insert using

    private void BulkInsert(DataTable dt)
    {
        string consString = _config.GetConnectionString("yourConnectionStringkey");
        using SqlConnection connection = new SqlConnection(consString);

        connection.Open();
        using var sqlBulkCopy = new SqlBulkCopy(connection, SqlBulkCopyOptions.TableLock | SqlBulkCopyOptions.UseInternalTransaction, null);
        sqlBulkCopy.DestinationTableName = "dbo.TargetDb";
        sqlBulkCopy.ColumnMappings.Add("Id", "Id");
        sqlBulkCopy.ColumnMappings.Add("Name", "Name");
        sqlBulkCopy.ColumnMappings.Add("NullableId", "NullableId");


        sqlBulkCopy.WriteToServer(dt);
        connection.Close();
    }

You dont have to do the column mappings given

    // Summary:
    //     Returns a collection of Microsoft.Data.SqlClient.SqlBulkCopyColumnMapping items.
    //     Column mappings define the relationships between columns in the data source and
    //     columns in the destination.
    //
    // Value:
    //     A collection of column mappings. By default, it is an empty collection.
    //
    // Remarks:
    //     If the data source and the destination table have the same number of columns,
    //     and the ordinal position of each source column within the data source matches
    //     the ordinal position of the corresponding destination column, the <xref:Microsoft.Data.SqlClient.SqlBulkCopy.ColumnMappings>
    //     collection is unnecessary. However, if the column counts differ, or the ordinal
    //     positions are not consistent, you must use <xref:Microsoft.Data.SqlClient.SqlBulkCopy.ColumnMappings>
    //     to make sure that data is copied into the correct columns. During the execution
    //     of a bulk copy operation, this collection can be accessed, but it cannot be changed.
    //     Any attempt to change it will throw an <xref:System.InvalidOperationException>.

These 2 lines lets you insert nullable values into the datatable column

 { ColumnName = property.Name, DataType = Nullable.GetUnderlyingType(property.PropertyType) ?? property.PropertyType, AllowDBNull = true }

newRow[property.Name] = item.GetType().GetProperty(property.Name)?.GetValue(item, null) ?? DBNull.Value;
Rightist answered 17/5, 2022 at 19:28 Comment(0)

© 2022 - 2024 — McMap. All rights reserved.