MartyConnelly
martyconnelly at shaw.ca
Fri May 11 18:24:18 CDT 2007
Go with a big 64 bit CPU box and Windows Vista 64 bit Business Version
that will give you access to 128 Gig of onboard RAM
JWColby wrote:
>Marty,
>
>My only question about this method is the loading of all the records at one
>go before processing. Remember that I am doing raw files that can be 4
>gigabytes of text, up to (in this case) 4 million records and (in this case)
>149 fields. These files are HUGE by desktop standards.
>
>My method uses a single line read / process / write and thus is pretty much
>guaranteed to handle any size file, any number of records, any number of
>fields.
>
>
>
>John W. Colby
>Colby Consulting
>www.ColbyConsulting.com
>
>-----Original Message-----
>From: dba-sqlserver-bounces at databaseadvisors.com
>[mailto:dba-sqlserver-bounces at databaseadvisors.com] On Behalf Of
>MartyConnelly
>Sent: Wednesday, May 09, 2007 2:04 PM
>To: dba-sqlserver at databaseadvisors.com
>Subject: Re: [dba-SQLServer] Bulk insert
>
>Another option is to use SqlBulkCopy a class that comes with Net Framework
>2.0 There is a SqlBulkCopy example in the book below that uses a CSV import.
>Ttricky to setup but it works. This new 2.0 class is designed to call the
>SQLSMO layer underneath the covers--that replaces SQL DMO.
>William Vaughn's Hitchhiker's Guide to Visual Studio and SQL Server (7th
>Edition)
>
>Here is C## code example of calling using a datareader stream into
>SQLBulkCopy This handles 40,000 rows a second, 8000 if you apply indexing
>
>private void button1_Click(object sender, EventArgs e) {
> Stopwatch sw = new Stopwatch();
> sw.Start();
> DataTable table = new DataTable();
> table.Columns.Add(new DataColumn("File",typeof(string)));
> table.Columns.Add(new DataColumn("IID",typeof(int)));
> table.Columns[1].AutoIncrement = true;
> table.Columns[1].AutoIncrementSeed = 1;
> table.Columns[1].AutoIncrementStep = 1;
>
> StreamReader sr = new StreamReader("c:\\filelist.txt");
> while (!sr.EndOfStream ){
> table.Rows.Add(sr.ReadLine());
> }
>
> sw.Stop();
> Debug.Write(sw.Elapsed.TotalSeconds + " seconds for " +
>table.Rows.Count + " = " + Convert.ToDouble(table.Rows.Count) /
>sw.Elapsed.TotalSeconds + "
>rows
>per second loaded to datatable");
> sw.Start();
>
> SqlConnection sqlcon = new SqlConnection("data
>source=lon0371xns;initial catalog=SonarBackup;integrated security=sspi");
> SqlBulkCopy bc = new SqlBulkCopy(sqlcon);
> bc.DestinationTableName = "FileList";
> bc.NotifyAfter = 5000;
> bc.SqlRowsCopied += new
>SqlRowsCopiedEventHandler(bc_SqlRowsCopied);
> bc.ColumnMappings.Add(new SqlBulkCopyColumnMapping("File", "File"));
> bc.ColumnMappings.Add(new SqlBulkCopyColumnMapping("IID", "IID"));
> sqlcon.Open();
> bc.BulkCopyTimeout = 500;
> bc.WriteToServer(table);
> sw.Stop();
>
> Debug.Write(sw.Elapsed.TotalSeconds + " seconds for " +
>table.Rows.Count + " = " + Convert.ToDouble(table.Rows.Count) /
>sw.Elapsed.TotalSeconds + "
>rows
>per second loaded to db");
>
>
>
>}
>
>
>
>JWColby wrote:
>
>
>
>>>The only caveat here is if you have empty fields in your file, a single
>>>
>>>
>>>
>>>
>>space is inserted instead of a null.
>>
>>What is it with all "the only caveat here" stuff? I am sure that there is
>>
>>
>a
>
>
>>darned good reason.
>>
>>In the end it is just easier to roll your own rather than work around the
>>issues that the built in stuff seems to have. I have 150 fields (in this
>>data set). Somehow I have to do an update on all 150 fields. I suppose I
>>could have my converter run 150 update queries to do each column. Or 700
>>update queries to do the next data set. Or just do the stripping of the
>>spaces external to SQL Server and be done with it. Either way I still have
>>to use my toy.
>>
>>Once I move up to VB.Net I will be able to use threads to do the stripping
>>and the BULK INSERT Sproc in parallel.
>>
>>BTW, I have to do something very similar all over again once I get the data
>>in. I will need to export the entire table back out, 2 million record sets
>>of data to delimited files for CAS / NCOA processing, dumping 100 million
>>records out into ~50 files (just the address data this time). The CAS /
>>NCOA process theoretically will process all files placed into an input
>>directory (input to that program), dumping the processed files into an
>>output directory (output from that program). At which point I have to pull
>>all of the CASS / NCOAd files BACK out of that output directory into to yet
>>another table. And that is just the "pre-processing".
>>
>>You might be getting a clue by now why I do not want to be manually doing
>>all the crapola involved with the solutions that do not involve an external
>>control process. Someday fairly soon I will have a completely automated
>>system for doing all this. I will be back to blowing bubbles and poking at
>>Charlotte with a big stick.
>>
>>John W. Colby
>>Colby Consulting
>>www.ColbyConsulting.com
>>
>>-----Original Message-----
>>From: dba-sqlserver-bounces at databaseadvisors.com
>>[mailto:dba-sqlserver-bounces at databaseadvisors.com] On Behalf Of
>>MartyConnelly
>>Sent: Tuesday, May 08, 2007 10:27 PM
>>To: dba-sqlserver at databaseadvisors.com
>>Subject: Re: [dba-SQLServer] Bulk insert
>>
>>Uhh, there is a one line fix to remove trailing blanks in SQL, different
>>defaults for SQL Server versions and NChar and VChar.
>>
>>SET ANSI_PADDING OFF
>>
>>When a table is created with the setting turned on (the default), spaces
>>are not trimmed when data is inserted into that table. When ANSI_PADDING is
>>off, the spaces are trimmed.
>>
>>So if you SET ANSI_PADDING OFF, create your table, then set it back on
>>again, when you bcp the data into the table, the excess trailing spaces
>>
>>
>will
>
>
>>be eliminated. The only caveat here is if you have empty fields in your
>>file, a single space is inserted instead of a null. If this is the case
>>
>>
>with
>
>
>>your data file, you will need to do an update to set columns to null when
>>len(yourcolumn) = 0.
>>
>>See BOL
>>http://msdn2.microsoft.com/en-us/library/ms188340.aspx
>>
>>http://msdn2.microsoft.com/en-us/library/ms187403.aspx
>>
>>
>>
--
Marty Connelly
Victoria, B.C.
Canada