Ask MetaFilter questions tagged with MSSQL

MS SQL please speed up so i dont cry

antonymous — Tue, 20 Oct 2009 09:07:43 -0800

MS SQL Pros, I need your help! What's causing MS SQL 2005 to import records so slowly? It's a persistent problem across all databases that just developed yesterday. Details inside... So my 3rd party app that I use to import records into MS SQL 2005 Standard has gone from approximately 10K records per hour to 500 records per hour. I've tried creating new databases to test on, and they all give me the same result.

I'm no MS SQL expert, but I did a bit of googling and created a trace file. It looks like the delay (according to the Start and End Time in the trace record) is between an "IF @@TRANCOUNT > 0 COMMIT TRAN" and "SELECT ID FROM tlbFolders WHERE NAME="C:\PATHTODATA\". The delay between these 2 events is about 10 seconds, and I can't tell why it's acting so slow.

I'll follow up in the thread, let me know if you need more info or know of any additional resources that can help me answer this question, or any good troubleshooting tips. Thanks!

Building Oracle support into a Microsoft-only product

relucent — Fri, 05 Jun 2009 09:37:55 -0800

Currently Microsoft-only, but wanting to change our software soon to support Oracle as well. What are we getting ourselves into? My company develops a web-based human resources management system and software suite written in .NET which uses a MS SQL backend. We have been a "Microsoft shop" for at least ten years, with all of our products being written in a Microsoft language using Microsoft tools on a Microsoft database.

We recently got a fairly sizable RFP from a company who is extremely interested in our software. The RFP describes our software almost perfectly and they really want to go with us — however, their one stipulation is that our software needs to support Oracle as a backend.

The dollar amount of the contract has the upper level management willing to do whatever it takes to secure this contract, but I'm worried that we're biting off a lot more than we can chew. No one in the whole company has any Oracle experience except for one developer who has said that it's a bad idea. We also have a very tight development schedule for the next year. Also, this is the first RFP we've gotten in many years which requires Oracle as a backend, so I don't think this change will help us earn more business in the future.

Knowing what you do about the differences between MS SQL and Oracle, how reasonable would it be for a small software company to build Oracle support into their flagship product, given limited development resources and almost no Oracle knowledge?

Upsize Access database?

cbrody — Tue, 10 Mar 2009 13:39:02 -0800

I have been asked to make a proposal to upsize an Access database to allow for a few more concurrent users (currently 10) and ideally improve the overall performance of the system. Am I getting in over my head? To give an idea as to the size of the project the current system has 35 tables, and on the order of 1000 queries and 1500 macros defined. Not even the original developer (the client) knows which ones are actually in use. There are on the order of 200,000 records in the current live system. The client is not willing to spend the money to have the system completely re-built (which I would estimate at being at 3-6 months’ work) but they would like some input as to how the current system could be improved, and they are probably willing to spend enough to cover two to three weeks of my time.

I have done some tests, successfully upsizing the tables to MySQL, but even a single insert from one of the simpler entry forms fails. I figure that with my basic knowledge of SQL that I could probably get a lot of it working (if not very efficiently), but would MS SQL be a better bet for compatibility and performance issues? I have a fair amount of experience with Access, having built customer management and billing systems from scratch, but am no SQL/RDBMS guru. I could use this work, but am I doomed to fail?

How can I import a 46GB database into MSSQL without ending up with a 46GB transaction log?

odinsdream — Thu, 29 May 2008 12:13:57 -0800

I need to import a database into MS SQL but I do not want to end up with a transaction log as large as the database. What's the best way to do the import? We're working with a software vendor to migrate their database into MS SQL. They set up an ODBC connection to their database and then use Enterprise Manager Data Transformation Services (DTS) packages to copy the tables into an MS SQL database.

However, at the end of their test pass we had a 46GB database file and a 46GB transaction log.

What's the best way to do this without creating such a giant log?

LAMP to WISA

gringoese — Sun, 18 May 2008 18:18:53 -0800

After working in a mostly LAMP (Linux, Apache, MySQL, PHP) environment the last few years, I need to make a quick switch to WISA (Windows, IIS, MS SQL, ASP). I have very little experience with WISA. Give me some tips to make a smooth transition. What is different and what is the same?

It keeps growing and growing and growing...

eas98 — Fri, 28 Mar 2008 09:42:35 -0800

Help me understand MSSQL logs... At least enough to do what I need to do. Ok, I know the logs are there for disaster recovery and such. I understand the concept of Point-in-Time restore. But what I don't get is why the logs don't get truncated when a backup of the DB is done. One would think that you would get a fresh new log started after a backup, but yet my log keeps growing and growing and growing.

Anyway, I'm sure there are reasons for it, but what I need is to be able to do backups where the log gets reduced in size when the backup is done, since I will no longer need the past logged data after that point.

For reference: I'm doing 'Full' recovery model, and a 'Full' (Backup Type) backup is done once daily, at night. I want to keep the logs for Point in Time restores for crashes that happen during the day. I don't need a log for what happened in prior days. The code used is like as follows:

BACKUP DATABASE [my_db]
TO DISK = @sBackupName
WITH INIT, NOUNLOAD, NAME = N'backup label', NOSKIP, STATS=10, NOFORMAT

Yes, I'm a SQL n00b.

Thanks for any help/info on this.

But I don't want to learn .NET!

logic vs love — Thu, 20 Mar 2008 16:01:23 -0800

How can I connect to a SQL Server 2005 database using windows authentication with PHP? I'm working on a small PHP/MSSQL application at work, but I can only connect to the database server using windows authentication and I can't figure out how to make that happen. Is it even possible?

Authentication across Windows, Linux, and Mac

remi — Fri, 11 Jan 2008 06:22:57 -0800

How do you use a single authentication system for different kind of servers, systems, and workstations? We're a small Web development business of less than 20 employees. Currently, we have three local servers:

Server A, using Windows Server 2003, is the "main" server for putting all of our document (including financial data) and most of our projects. For development and testing, we also host Web sites there which require a Windows server, ASP and MS SQL. It's also our main DHCP and DNS server.

Server B, Fedora Core 6, is mainly for hosting projects requiring Linux, PHP or ColdFusion, and MySQL. It's the backup DHCP and DNS server.

Server C, Windows Sever 2003 Web Edition, is for hosting projects requiring ColdFusion and Windows.

We have various kinds of workstations at the office: Windows XP, Windows Vista, Ubuntu, Fedora, and Mac OS X 10.4.

Our biggest annoyance at the moment is authentication. Every employee has a different account for everything. Windows file shares, Linux file shares, Linux shell accounts, MySQL, MS SQL...

Server A already has Active Directory set up, though I'm not too familiar with it. (I'm more of a Linux system administrator.) I tried some ways to combine all one employee's accounts together, but it just won't work. One of the ways I tried was to set up PAM on server B to use LDAP or other mechanisms supposedly supported by the server A, but it doesn't work.

Now, I'm wondering, are there any methods to make the authentication process of every service work together?

I'm not even sure where to start to set this all up. Any suggestions will be appreciated! Thanks in advance!

Right Place, Wrong Time

dw — Tue, 13 Nov 2007 18:37:07 -0800

I have a set of dates/times in a MS SQL table that lack a timezone offset. I want to convert them to UTC, but I have to be mindful of Daylight Time for each of these dates (they're in the future). How can I convert them to UTC correctly? The nuts and bolts are these:
We have a calendaring application with an MS SQL Server 2000 backend. When event dates get entered into the system, they're entered without a timezone code, e.g.:

2007-11-13 12:00:00.000

Each of these events is in the future. Some are in March or April. We observe DST.

I'm making changes to the data feed so that we can implement the hCalendar microformat on our calendars. The problem is the time zone. hCalendar calls for dates like:

2007-11-13T12:00:00-08:00

Originally, one of the engineers slapped the current time offset from UTC on there. Unfortunately, if you're looking at the calendar while we're on standard time, an event that will be in DST will not be offset correctly, being an hour earlier or later.

So, back to the drawing board. We decided to convert all the dates to UTC before putting them in the data feed for the microformat. But the time zone problem came right back.

Is there an easy way to convert these future dates to the correct UTC that's mindful of what the DST offset is in the future? There's no function in MS SQL 2000 that allows you to look up whether a date is PDT or PST, is there?

Where's Enterprise Manager?

SpecialK — Wed, 01 Aug 2007 15:27:31 -0800

I have Visual Studio 2005 Professional. I installed VS2k5Pro with all of the options except the sample databases on my WinXP box, but now I can't figure out how to access the Enterprise Manager - replacement ... Enterprise Management Studio, or even find out if that app is included with the version I have. Any suggestions?

Mapping drive to MSSQL table

seanyboy — Tue, 30 Jan 2007 07:08:31 -0800

Is there a way of mapping a Virtual Drive letter or a share to an MSSQL database? In windows, Is it possible to create a mapped drive or a shared folder, where the files are actually held in an MSSQL (or equivalant) database? i.e. If I save a file called "new file.txt" on Z:\, it actually saves the file as a new row in an SQL table.

Host-Filter

strawberryviagra — Fri, 27 Oct 2006 19:08:48 -0800

Looking for a dedicated server hosting provider We’re about to launch an online application – built in Flash to run research experiments.

The parameters that are most important are:

- Low latency (we’ll be running experiments in all parts of the World including China),
- High spec box (probably looking at 200 concurrent users hitting an app with a lot of optimised animations and rich media, regularly writing to a database),
- Budget – preferably less than $800USD per month.

*It has to be a Windows Server running >IIS5 with MS SQL database (and have SQL server enterprise manager)
Any other suggestions appreciated (except technology choice – it’s too late and complicated to get into those types of fundamentals).

A query connundrum.

Civil_Disobedient — Sun, 03 Sep 2006 05:51:38 -0800

Attention SQL gurus. This has me completely stumped. How can I find the index (position) of a row in a table without querying the entire set? Here is the situation:

I have a web application that has some ajaxified tables on many of the pages. These tables show various slices of data. Most of the tables are about 10 rows per page, so for example, if the table was a list of all Customers, and you've got 10,000 customers, that would be 1,000 pages of customers.

To initially show the table, I have to do two queries: one is the initial 10 rows that are to be displayed, the other is a count(*) to get the total amount of rows.

Here is the problem:

I have a particular customer ID in my hand. I'd like the webtable to take this ID and jump to the correct page that will show that customer. Naturally, it has to take into account sorting.

Hypothetical simplified example:

ID    NAME

1102  C_D's Tavern

1294  Bob's Discount Dairy

1299  Chico's House of Wicker

...

9309  Larry's Llama Emporium

If I sort by ID (natural order), the index of Bob's Discount Dairy would be 2, Chico's Hosue of Wicker would be 3, etc. If I'm displaying "page 1", that means the first 10 indexes, so I know those two would be on the first page. "OK, so what page is ID# 5129 on?" That's the problem.

What I want to know is if there's a way to figure out the index of a particular ID without having to query the entire dataset. Or if I do have to query the entire dataset, I'd like it to be as painless as possible. Another important note: some of these queries are being run off of VIEWs, which could complicate things.

This is on Microsoft SQL Server 2000.