There are a number of articles (like this one) explaining how to do this for SQL Server 2000, but it’s so much easier in SQL Server 2005 that I’ll post the 2005-compatible code here. You should read this Microsoft Research article to get the Microsoft best practices justification for storing various file sizes in the database versus on the filesystem. Note, the following code sample is extremely simplistic and various iterations of similar logic are available from Microsoft and others.

if exists(select * from sysobjects where type=N'U' and name=N'picTable')
drop table dbo.picTable;
go
create table dbo.picTable(FileName sysname, Document image)
go
if exists(select * from sysobjects where type=N'P' and name=N'spLoadImagesFromLocalFolder')
drop procedure dbo.spLoadImagesFromLocalFolder;
go
create procedure [dbo].[spLoadImagesFromLocalFolder](@directory sysname,@fileextention nvarchar(10))
as
begin
declare @cmd varchar(8000),@fname nvarchar(200),@sqlstmt nvarchar(500),@ParmDefinition nvarchar(500),@cnt int

if object_id(’tempdb..#filestoload’) is not null
begin
drop table #filestoload
end
create table #filestoload ( RowNum int identity, fname varchar(8000) null)

SET @cmd = ‘dir ‘ + QUOTENAME(@directory,’”‘) + ‘ /T:W /A:-D /b’
insert into #filestoload ([fname])
exec [master].[dbo].[xp_cmdshell] @cmd
/* print @cmd */
delete from #filestoload where fname is null or lower(fname) not like N’%.’+lower(@fileextention)
select @fname = ”

select @cnt=count(*) from #filestoload
set @ParmDefinition = N’@fname nvarchar(200)’;
while @cnt>0
begin
select @fname = fname from #filestoload where RowNum > @cnt
set @sqlstmt=N’INSERT INTO picTable(FileName, Document) SELECT ”’+@fname+N”’ AS FileName,* FROM OPENROWSET(BULK N”’+@directory+@fname+N”’, SINGLE_BLOB) AS Document’
/* print @sqlstmt */
exec sp_executesql @sqlstmt,@ParmDefinition,@fname=@fname
select @cnt=@cnt-1;
end
drop table #filestoload;
end
go
exec [dbo].[spLoadImagesFromLocalFolder] N’C:\’,N’JPG’

Hope that helps!