We Keep Coding

Xml data file to Sql Database including data comparison and validation

Evening all,
Background: I have many xml data files which I need to import in to an SQL database via my C# WPF based Windows Application. Whilst some data files are a simple and straight forward INSERT, many require validation and verification checks, require GUIDs from existing records (from within the db) before the INSERT takes place in other to maintain certain relationships.
So I've broken the process in to three stages:-
Validate records. Many checks exist such as e.g. 50,000 accounts in xml file must have a matching reference with an associated record already in the database. If there is no corresponding account, abandon the entire import process; Another would be, only the 'Current' record can be updated, so again, if it points to a 'historic' record then it needs to crash and burn.
UPDATE the associated database records e.g Set from 'Current' to 'Historic' as record is to be superseded by what's to come next...
INSERT the records across multiple tables. e.g. 50,000 accounts to be inserted.
So, in summary, I need to validate records before any changes can be made to the database with a few validation checks. At which point I will change various tables for existing records before inserting the 'latest' records.
My first attempt to resolve this situation was to load the xml file in to XmlDocument or XDocument instance, iterate over every account in the xml file and perform an sql command for every account. Verify it exists, verify its a current account, change the record before inserting it. Rinse and repeat for thousands of records - Not ideal to say the least.
So my second attempt is to load the xml file in to a data table. Export the corresponding accounts from the database in to another data table and perform a nested loop validation e.g. does DT1.AccountID exist in DT2.AccountID, move DT2.GUID to DT1.GUID etc etc. I appreciate this could also be a slow process. That said, I do have the luxury then of performing both the UPDATE and INSERT stored procedures with a table value parameter (TVP) and making use of the data table information.
I appreciate many will suggest letting the SQL do all of the work but i'm lacking in that skillset unfortunately (happy to learn if thats the general consensus) but I would much rather do the work in C# code if at all possible.
Any views on this are greatly appreciated. Many of the questions I've found are around bulk INSERT, not so much about validating existing records, following by updating records followed by inserting records. I suppose my question is around the first part, the validation. Does extracting the data from the db in to a data table to work on seem wrong, old fashioned, pointless?
I'm sure i've missed out some vital piece of information so apologies if unclear.
Cheers

C# and SQL Server 2008 - Batch Update

I have a requirement where I need to update thousands of records in live database table, and although there are many columns in this table, I only need to update 2-3 columns.
Further I can't hit database for thousand times just for updating which can be done in a batch update using SQL Server Table Valued Parameter. But again I shouldn't update all thousands records in one go for better error handling, instead want to update records in batches of x*100.
So, below is my approach, please give your valuable inputs for any other alternatives or any change in the proposed process -
1 Fetch required records from database to List<T> MainCollection
2 Save this collection to XML file with each element Status = Pending
3 Take first 'n' elements from XML file with Status = Pending and add them to new List<T> SubsetCollection
4 Loop over List<T> SubsetCollection - make required changes to T
5 Convert List<T> SubsetCollection to DataTable
6 Call Update Stored Procedure and pass above DataTable as TVP
7 Update Status = Processed for XML Elements corresponding to List<T> SubsetCollection
8 If more records with Pending status exists in XML file, go to Step# 3.
Please guide for a better approach or any enhancement in above process.

I would do a database-only approach if possible and if not possible, eliminate the parts that will be the slowest. If you are unable to do all the work in a stored procedure, then retrieve all the records and make changes.
The next step is to write the changes to a staging table with SQL Bulk Copy. This is a fast bulk loaded that will copy thousands of records in seconds. You will store the primary key and the columns to be updated as well as a batch number. The batch number is assigned to each batch of records, therefore allowing another batch to be loaded without conflicting with the first batch.
Use a stored procedure on the server to process the records in batches of 100 or 1000 depending on performance. Pass the batch number to the stored procedure.
We use such a method to load and update millions of records in batches. The best speed is obtained by eliminating the network and allowing the database server to handle the bulk of the work.
I hope this might provide you with an alternate solution to evaluate.

It may not be the best practice but you could embed some logic inside a SQL Server CLR function. This function could be called by a Query,StoProc or a schedule to run at a certain time.
The only issue I can see is getting step 4 to make the required changes on T. Embedding that logic into the database could be detrimental to maintenance, but this is no different to people who embed massive amounts of business logic into StoProcs.
Either way SQL Server CLR functions may be the way to go. You can create them in Visual Studio 2008, 2010 (Check the database new project types).
Tutorial : http://msdn.microsoft.com/en-us/library/w2kae45k(v=vs.80).aspx

Performance issues with transpose and insert large, variable column data files into SQL Server

I'm currently working on a project where we have a large data warehouse which imports several GB of data on a daily basis from a number of different sources. We have a lot of files with different formats and structures all being imported into a couple of base tables which we then transpose/pivot through stored procs. This part works fine. The initial import however, is awfully slow.
We can't use SSIS File Connection Managers as the columns can be totally different from file to file so we have a custom object model in C# which transposes rows and columns of data into two base tables; one for column names, and another for the actual data in each cell, which is related to a record in the attribute table.
Example - Data Files:
Example - DB tables:
The SQL insert is performed currently by looping through all the data rows and appending the values to a SQL string. This constructs a large dynamic string which is then executed at the end via SqlCommand.
The problem is, even running in a 1MB file takes about a minute, so when it comes to large files (200MB etc) it takes hours to process a single file. I'm looking for suggestions as to other ways to approach the insert that will improve performance and speed up the process.
There are a few things I can do with the structure of the loop to cut down on the string size and number of SQL commands present in the string but ideally I'm looking for a cleaner, more robust approach. Apologies if I haven't explained myself well, I'll try and provide more detail if required.
Any ideas on how to speed up this process?

The dynamic string is going to be SLOW. Each SQLCommand is a separate call to the database. You are much better off streaming the output as a bulk insertion operation.
I understand that all your files are different formats, so you are having to parse and unpivot in code to get it into your EAV database form.
However, because the output is in a consistent schema you would be better off either using separate connection managers and the built-in unpivot operator, or in a script task adding multiple rows to the data flow in the common output (just like you are currently doing in building your SQL INSERT...INSERT...INSERT for each input row) and then letting it all stream into a destination.
i.e. Read your data and in the script source, assign the FileID, RowId, AttributeName and Value to multiple rows (so this is doing the unpivot in code, but instead of generating a varying number of inserts, you are just inserting a varying number of rows into the dataflow based on the input row).
Then pass that through a lookup to get from AttributeName to AttributeID (erroring the rows with invalid attributes).
Stream straight into an OLEDB destination, and it should be a lot quicker.

One thought - are you repeatedly going back to the database to find the appropriate attribute value? If so, switching the repeated queries to a query against a recordset that you keep at the clientside will speed things up enormously.
This is something I have done before - 4 reference tables involved. Creating a local recordset and filtering that as appropriate caused a speed up of a process from 2.5 hours to about 3 minutes.

Why not store whatever reference tables are needed within each database and perform all lookups on the database end? Or it may even be better to pass a table type into each database where keys are needed, store all reference data in one central database and then perform your lookups there.

Faster way to update 250k rows with SQL

I need to update about 250k rows on a table and each field to update will have a different value depending on the row itself (not calculated based on the row id or the key but externally).
I tried with a parametrized query but it turns out to be slow (I still can try with a table-value parameter, SqlDbType.Structured, in SQL Server 2008, but I'd like to have a general way to do it on several databases including MySql, Oracle and Firebird).
Making a huge concat of individual updates is also slow (BUT about 2 times faster than making thousands of individual calls (roundtrips!) using parametrized queries)
What about creating a temp table and running an update joining my table and the tmp one? Will it work faster?

How slow is "slow"?
The main problem with this is that it would create an enormous entry in the database's log file (in case there's a power failure half-way through the update, the database needs to log each action so that it can rollback in the event of failure). This is most likely where the "slowness" is coming from, more than anything else (though obviously with such a large number of rows, there are other ways to make the thing inefficient [e.g. doing one DB roundtrip per update would be unbearably slow], I'm just saying once you eliminate the obvious things, you'll still find it's pretty slow).
There's a few ways you can do it more efficiently. One would be to do the update in chunks, 1,000 rows at a time, say. That way, the database writes lots of small log entries, rather than one really huge one.
Another way would be to turn off - or turn "down" - the database's logging for the duration of the update. In SQL Server, for example, you can set the Recovery Model to "simple" or "bulk update" which would speed it up considerably (with the caveat that you are more at risk if there's a power failure or something during the update).
Edit Just to expand a little more, probably the most efficient way to actually execute the queries in the first place would be to do a BULK INSERT of all the new rows into a temporary table, and then do a single UPDATE of the existing table from that (or to do the UPDATE in chunks of 1,000 as I said above). Most of my answer was addressing the problem once you've implemented it like that: you'll still find it's pretty slow...

call a stored procedure if possible

If the columns updated are part of indexes you could
drop these indexes
do the update
re-create the indexes.
If you need these indexes to retrieve the data, well, it doesn't help.

You should use the SqlBulkCopy with the KeepIdentities flag set.
As part of a SqlTransaction do a query to SELECT all the records that need updating and then DELETE THEM, returning those selected (and now removed) records. Read them into C# in a single batch. Update the records on the C# side in memory, now that you've narrowed the selection and then SqlBulkCopy those updated records back, keys and all. And don't forget to commit the transaction. It's more work, but it's very fast.

Here's what I would do:
Retrieve the entire table, that is, the columns you need in order to calculate/retrieve/find/produce the changes externally
Calculate/produce those changes
Run a bulk insert to a temporary table, uploading the information you need server-side in order to do the changes. This would require the key information + new values for all the rows you intend to change.
Run SQL on the server to copy new values from the temporary table into the production table.
Pros:
Running the final step server-side is faster than running tons and tons of individual SQL, so you're going to lock the table in question for a shorter time
Bulk insert like this is fast
Cons:
Requires extra space in your database for the temporary table
Produces more log data, logging both the bulk insert and the changes to the production table

Here are things that can make your updates slow:
executing updates one by one through parametrized query
solution: do update in one statement
large transaction creates big log entry
see codeka's answer
updating indexes (RDBMS will update index after each row. If you change indexed column, it could be very costly on large table)
if you can, drop indices before update and recreate them after
updating field that has foreign key constraint - for each inserted record RDBMS will go and look for appropriate key
if you can, disable foreign key constraints before update and enable them after update
triggers and row level checks
if you can, disable triggers before update and enable them after

Application aware data import

I'm building an application to import data into a sql server 2008 Express db.
This database is being used by an application that is currently in production.
The data that needs to be imported comes from various sources, mostly excel sheets and xml files.
The database has the following tables:
tools
powertools
strikingtools
owners
Each row, or xml tag in the source files has information about 1 tool:
name, tooltype, weight, wattage, owner, material, etc...
Each of these rows has the name of the tool's owner this name has to be inserted into the owners table but only if the name isn't already in there.
For each of these rows a new row needs to be inserted in the tools table.
The tools table has a field owner_id with a foreign key to the owners table where the primary key of the corresponding row in the owners table needs to be set
Depending on the tooltype a new row must be created in either the powertools table or the strikingtools table. These 2 tables also have a tool_id field with a foreign key to the tools table that must be filled in.
The tools table has a tool_owner_id field with a foreign key to the owners table that must be filled in.
If any of the rows in the importfile fails to import for some reason, the entire import needs to be rolled back
Currently I'm using a dataset to do this but for some large files (over 200.000 tools) this requires quite a lot of memory. Can anybody think of a better aproach for this?

There are two main issues to be solved:
Parsing the a large XML document efficiently.
Adding a large amount of records to the database.
XML Parsing
Although the DataSet approach works, the whole XML document is loaded into memory. To improve the efficiency of working with large XML documents you might want look at the XmlReader class. The API is slightly more difficult to use than what DataSet provides. But you will get the benefit of not loading the whole DOM into memory at once.
Inserting records to the DB
To satisfy your Atomicity requirement you can use a single database transaction but the large number of records you are dealing with for a single transaction is not ideal. You will most likely incur issues like:
Database having to deal with a large number of locks
Database locks that might escalate from row locks to page locks and even table locks.
Concurrent use of the database will be severely affect during the import.
I would recommend the following instead of a single DB transaction:
See if it possible to create smaller transaction batches. Maybe 100 records at a time. Perhaps it is possible to logically load sections of the XML file together, where it would be acceptable load a subset of the data as a unit into the system.
Validate as much of your data upfront. E.g. Check that required fields are filled or that FK's are correct.
Make the upload repeatable. Skip over existing data.
Provide a manual undo strategy. I know this is easier said than done, but might even be required as an additional business rule. For example the upload was successful but someone realises a couple of hours later that the wrong file was uploaded.
It might be useful to upload your data to a initial staging area in your DB to perform validations and to mark which records have been processed.

Use SSIS, and create and ETL package.

Use Transactions for the roll back feature, and stored procedure that handle creating/checking the foreign keys.