Slashdot Mirror
How Do I Make Sense of Microsoft Access?
Anthony Boyd asks: "I have a pretty good tool-set for LAMP work, but as I get into Microsoft jobs, I've started to wonder if I'm working with the best tools. In particular, I'm exploring an 'out of control' Microsoft Access setup, which has about 200 tables in 30 .mdb files, including some duplicated/outdated tables. I'd like to print the properties of each table (with the comments for each field), print the table list for each database, get info on the field types & relationships, and so forth. What tools do you suggest for trying to grok a large Access mess?"
Look for the Database Documenter in the tools menu...it will print out basically all of the information that you say you're looking for...
I'm going to assume you've got Access 2003, because I can't remember what they changed from 2000 anymore :-)
The Microsoft Access Conversion Toolkit will give you some of the information you want, and can be used to query MDBs network-wide. If you just need to figure out the mess of a single chosen database, start using the built-in features of Access. Check the relationships and see if anything has been diagrammed out for you. If it has, then you have the ERD ready. If not, have fun figuring it out... Use the stuff in 'Tools->Analyze' to get more property and design information. Try right-clicking on a table/query/form/report and selecting 'Object Dependencies'. This will allow you to see what requires it to work, as well as what the object depends on. Lastly, I just start working through the code/macros (yuck). The object dependencies stuff won't check macros or VBA, so you have to check manually. Sometimes you'll find DAO/ADO code opening connections programmatically.
Best of luck to you! This will suck badly, in case you didn't figure that out already. Access provides an upsizing wizard that can help you upload your data to an MS-SQL server, but that will require you debug (ADP as frontend) / rewrite (VB.NET) the forms and stuff.
Despite what people say, Access does allow for security rights. However, it is not linked in any way to the machine or Active Directory. You use a modified shortcut to load the database with a security file. It works alright for most things, but there is no record-level security, and it sucks when you have 20 people signed in and you have to update the file. Also, supposedly there are cracks that break that security.
This all leads me to my next point for all who read: DO NOT USE ACCESS AS AN ENTERPRISE-LEVEL / MISSION CRITICAL DATABASE SUITE. Pay for a decent tool/programmer/dba/whatever if you really like your data. This application is just for personal / small-group data storage. There is a reason it comes with Office, and not SQL Server. Thank you.
And if not LAMP, then atleast something based on MSSQL. Access is not meant to handle things with that many tables and presumably quite a bit of (important) data. Not to mention security and reliability.
Every time you post an article on Slashdot, I kill a server. Think of the servers!
but it does the job.
On the menu, pull up Tools, Analyzer, Documenter, and pick your criteria. Access is clumsy and not secure, but it's also what most organizations have. That doesn't keep the rest of us from attempting to subvert from within...
To consolidate the duplicate tables, build a query that replicates grep and/or another that replicates diff, and have fun from there. Somehow I'm sure that you know how to do this.
Even though I prefer to work with other platforms and venues, my Access skills have managed to keep me employed and the cats fed while I decide what to do next with my life.
I second TOAD, having used it on a daily basis with an absolutely massive (hundreds of tables, many millions of lines) Oracle-based system, and it's been the best way I've found of making sense of things. The "Schema Browser" function I find particularly helpful when I know vaguely what column I'm looking for, but not what table it's in. It's replaced a lot of the old "cheat sheets" I used to have pinned to every flat vertical surface in my cube.
I've heard it's a fairly expensive piece of software, but thankfully I don't pay for it. It might be tough to get your PHB to spring for it, if that's actually the case...but I've yet to use or even hear of a better way to work with really complex DB systems.
"Ladies and gentlemen, my killbot features Lotus Notes and a machine gun. It is the finest available."
There are a ton of commercial utilities and add-ons for MS Access (check out Access Advisor at your local bookstore), but most of those are just VB apps or ActiveX controls that just do what you could do yourself with a little Visual Basic. Once you have got the basics down from some online tutorials, Access Cookbook by Kurt Getz is a great investment.
.NET.
MS Access has a large community online, especially comp.databases.ms-access. Google is your friend - just about everything you'll ever want to do has already been done and has VB code examples online.
Here is a thread that has code demonstrating how to dump the contents of an Access database as DDL into text files:
comp.databases.ms-access: Exporting jet table metadata as text?
PS - If you are impatient with the limitations of VBA (aka "VB Classic"), there are Microsoft Office interop libraries that will let you automate Access Databases in
Try Tora. It's a great database admin tool. There's a couple things Toad will do that Tora won't, but I've found it to be a nice substitute.
Free
Cross Platform
Open Source
supports Multiple Databases
Really quite a nice application.
Access has it's place. Ignore the OS zealots here, in the real world Access is a viable solution in many business contexts and there is no direct OS equivalent that comes anywhere close, besides if you try to convert it all to LAMP you'll most probably go out of business in the time it will take you :-).
:-). Both of these are time-consuming to replace by a browser interface: the Forms will need extensive use of AJAX to reproduce the immediate responsive feel of an Access application, and Reports can take a lot of coding and even then reproducing pagenation and so forth can be problematic.
OK, firstly you need to convert the data from Access to SQL Server. This is essential: Access is really an excellent front-end system, but it's data handling sucks big time in a mult-user situation. The upsizing wizards included with it do a pretty good job, although you'll undoutably want to tweak manually. Upsizing then reviewing the database and adding relational integrity and other database rules is an essential first step. If you can't afford SQL Server, which is probably on of MS's best products, the MSDE will do if you don't have too many users. It's also a little known fact that Access will run as a front end to Postgres - although I've never tried this myself if you Google for it there's quite a few resources out there.
Having upsized the database you then have two choices. Review and modify the upsized MDB front-end, or create a new ADP project and convert. ADP's have some advantages, but you have to convert manually and rewrite the data access code. This does take time (although a suprising amount can be cut and pasted from between MDB and ADP). The choice here is heavily influence by how the old database has been written - I've seen some Access applications which are practically VBA applications and need to be rewritten susbtantially to use SQL Server even if left as MDBs, whereas others hardly need any changes at all. If you are picking up the application to support yourself and can afford the time the ADP approach is probably prefable as you'll get to know the code and iron out any junk.
You should indeed consider what use you can make of LAMP (or more accuratly WISP). Access's strengths lie in it's ability to support detailed responsive Forms for data entry and most particularly it's Reports where complex output can be generated remarkably quickly - generally it's RAD abilities blow LAMP and similar away for anything but a simpler application (and Ruby on Rails, that includes you
However in a business context it's quite usual for there to be a core group of users who are responsible for data entry and 'expert' use of the system, and a wider group of users who need just read-only access or some very simple data entry, generally for a limited number of screens. If this is the case it's a viable strategy to replace Access by a browser interface for these users. PHP runs happily on a windows server so all your LAMP skills can be applied quite readily. THE major advantage of replacing Access for the casual users is that you then no longer need to deploy Access, which will save you both licence money and support time.
If at a later date you have the time and motivation to convert more of the core user functionality from Access to browser then you can do that. A viable strategy is to convert the Forms but leave the complex reports in a Access as a 'reporting suite'. In many business setups it's quite common to find an 'expert user' who is capable of creating bespoke reports in Access. Handled correctly these people can be a valuable asset - generally I create an 'Adhoc' or 'Scratch' Access application for them (mdb is strongly preferred in this case so that objects are not created on the server) which they can use to generate bespoke reports. The core functionality is placed in a separate Access application which they do not modify.
Take-home message is to recognise that all these technologies have strengths and weaknesses and play
If your bosses will shell out for it, then Visual Studio 2005's Integration Services can take data from any number of MDBs, Excel files, text files, databases, etc and transform them into whatever you want. We recently used it to move a creaking Access DB (1 table, 165 columns!!!) into a (temporary, and slightly more normalised) SQL Server schema - we had cursors running to generate keys, data cleansing procedures plugged in, and best of all it ran at the click of a button, so we could test it very easily with copies of the data before the final rollout.
"If he were a plant, people would roll him up and smoke him."
I'll agree with most of what Cruchan has to say, but want to add in a perspective that I've found important. Almost any business case that is capable of making good use of Access's capabilities is also capable of outgrowing Access's utility. Let me be more specific.
.mdb file and re-creating the data interface in HTML forms. The first is easy via an SQL dump, which Access supports. The second is more difficult because what you're essentially doing is Industrial Engineering.
Access's primary strength is that it allows a novice but intelligent user to store data in a database and create views with which they can examine, alter, and add to that data. This makes it very attractive for many small business owners to create methods of keeping track of customers, sales, products, whatever. If your business stays small, then you've saved yourself a bit of money and solved a paperwork headache.
If your business (venture, hobby, whatever) grows, then you will invariably run into Access's limitations. It's very easy to use, but a database program of any complexity will eventually run into programming errors that result in data corruption. They aren't everywhere, but when you run into one you're pretty well hosed. Microsoft may have fixed all of these kinds of bugs between their 2000 and 2003 release - I've somewhat gotten out of that kind of business - but I somehow doubt that Microsoft has changed its philosophy that much.
When you do run into that kind of problem, you have two choices. Keep a second paper trail of all of your changes so that you can fix the database when it hits that bug, or pay someone to migrate you to something more reliable. You would think that there would be an option to pay someone to fix your Access implementation, but by the time someone is willing to shell out money for this kind of thing you can pretty well guarantee that the flaws are in Microsoft's software, not in anything the user has done with it.
So on to the user's question - what do you do when you hit that tree and fly through your windshield? You have to remember that Accesss is a front end database management tool, not an actual database. What you need to replace isn't the actual storage of the information, but the routines that alter and display it. LAMP is an entirely viable idiom for this kind of change, even in a Windows environment. I run a LAMP environment on my laptop so I can develop and show off my web site designs while not online, and it's very reliable. Additionally, it allows remote access of information from many locations, although it takes a bit of skill to write something that can be altered from many locations at the same time.
I don't suggest WISP simply because any further growing will either lock you into Microsoft tools (many of which are highly suspect) or result in tedious and expensive searching for obscure features that allow you to attach other people's tools to the Microsoft Architecture.
Moving away from Access involves two things - migrating the data out of that
When you're talking about a 200 table database, you can quite readily start with the Database Documenter. It'll spit out a bunch of stuff that'll tell you what the formats of the tables are, but won't tell you how they're hooked together or what they're used for (unless the person who made the database was very, very professional). Then you need to have someone show you how the database is used. From that you can figure out what the inputs of the data process are, where the information is put, what is done with it, and what form and place it has to exist for it to be viewed and outputted.
From that you can generate charts that show how the various tables are connected, identify what the rules are (all cars shalt have a color), and get an idea for process flow.
No, this isn't easy. For really big databases it can take months. I was on a project where we were attempting to reconcile seven regional AAA databases into a central database and the entire project collapsed for the inability to get someone to tell us how the data was used. There just isn't enough information in a database dump to determine this.
Best of luck.
Wake up - the future is arriving faster than you think.
Until you get that POS replaced or moved to a real database, you can automate repair with a tool called jetcomp. It can repair many databases that Access's Compact and Repair and DAO's CompactDatabase method can't, and can be automated from the command line.
I went so far as to write a script to detect when the database has been corrupted by attempting to open it and checking the error message if it can't, and automatically backup and repair it using jetcomp. Our corruption problem finally went away when I found a system that had a slightly bad network connection and fixed it. Network problems become more apparent if you specify a larger packet size when pinging, like 400 bytes or so. We went from almost daily corruption to no corruption in over a month.
If the former programmers were "macro happy" it is very difficult to decode, but with time, a lot of paper to map it out, and even more patients, you can get it done.
Good luck with your project. If you need any help, reply to this message and we can talk.
Randy