If you’ve already developed COM APIs with .Net, typically in C# with VBA as the consuming language, you’ve probably leveraged two powerful features:
by-reference parameter passing that allows the API to change the input object itself, not only its content.
This is marked with ref in C#:
void LoadData(ref string data)
optional parameters that allow the caller to omit them when calling into the API:
void SaveData(string data = "DEFAULT")
Sometimes you want to have both for a parameter: the ability to omit it and to change it.
This is something COM supports but unfortunately this is not supported by C#.
Recently one of my customer needed precisely that for a COM API developed in C# migrated from VB6 where it was plug-and-play.
The original developers of the API had done a quick study and concluded that it was not possible with the new C# API and indeed it was not documented at all.
From experience I know that COM is not used a lot today and that resources are scarce, so, to be sure, I’ve done my due diligence and reexamined the issue, digging a little deeper and testing undocumented stuff.
I’ve finally found a solution that I would describe as a “trick”, because it’s not documented, and that I’ll present in detail in this article. Continue reading →
Note to pedantic guys: yes Access is not a database engine, only the graphical front-end to Jet/ACE, but we’ll stand with this simplification.
The sample application (Excel + VBA) and the SQL source code are available in this ZIP archive: VBA_Databases_Source.zip.
If you are an advanced VBA developer who builds applications that manage a non trivial amount of data odds are good you are using an Access database to store them.
If this setup perfectly fits your current needs, you completely master it, you’re not experiencing any issue and your needs won’t evolve in the near future you can skip this article and continue enjoying your system.
Indeed, do you really need a new database management system (DBMS)?
Often the only argument in favor of migrating to other DBMS is they are “better”; while it’s true for some technical capabilities, it may not be regarding other “metrics” like simplicity: Access is easy to understand and manage for non IT staff and is often installed with default business workstation setup along with the rest of the Office suite.
So let’s say you have strong reasons to migrate to a new DBMS because you’ve come to the point where you feel you need support for at least one of the following features: interoperability, big storage, high concurrency (hundreds of users) and/or high performance, and Access starts to become a part of the problem.
So what can you do if you want to enhance the quality of your database without making your total cost of ownership (TCO) explode?
Your TCO is essentially made of:
licensing costs: limiting them is quite simple: using a free, usually open-source, database and paying only for support
management costs: they are by far bigger than the licensing costs and are directly impacted by the complexity of the DBMS; so you need a simple DBMS that you’ll be able to setup and manage yourself as you used to do with Access without the help of system or database administrators
development costs: every additional change to your current schema or VBA implementation to fit the new DBMS will have a cost; so we want things to be transparent with zero additional development, which in particular means a standard SQL-based DBMS.
While the equation may seem a little complex it has at least three solutions:
SQLite is the ideal candidate if you’re happy with the “single-file” model, you don’t have high concurrency constraints, and your only needs are interoperability (with Mac OS, Linux, Unix…), bigger storage and/or costs savings,
MySQL and PostgreSQL: if you need support for high-concurrency, really big storage (e.g. tens of GBs), advanced user-management, performance fine tuning and other advanced features you’ll have to jump out of the single-file world.
If you don’t have specific requirements then MySQL and PostgreSQL will appear similar to you and equally do the job. However, in this kind of situation, I have a preference for MySQL, not because its inherent capabilities would be superior (as I’ve said MySQL and PostgreSQL are roughly equivalent for simple setups), but because, as the reference open-source DBMS for years, MySQL benefits from a huge community and toolbox. Moreover, while you’ll sure find the tools to work in good conditions with PostgreSQL, if you ever need to justify your choice to your hierarchy you’ll be in a better position if you choose the standard solution instead of the challenger.
But as I’m not sectarian, and for completeness, I’ll cover both.
In this article I’ll quickly cover the setup of these three DBMS (with links to other resources for more extensive instructions) and illustrate their usage with a small VBA application, a revolutionary todo-list manager, that uses Access too.
If you have an important VBA code base you know how difficult it is to maintain it, not because VBA is inherently a bad or poor platform but because the developers are often either end-users not aware of good development practices or professional developers not skilled enough in VBA development who have learnt it on the job. In both cases, more often than not, you end up with poor quality source code.
There is other forces that make VBA development less appealing, as the pressure of software vendors like Microsoft who’d like to sell you not just Office but Office+Visual Studio and then promotes other technologies like .Net with languages such as C# and VB.Net. Just have a look at the evolution of the VBA editor design and capabilities since 10 years and you’ll understand that it does not benefit from fanatic development and promotion efforts.
It’s why you should avoid the headlong rush and restrict your VBA development efforts to the bare minimum: for new developments you should consider other languages and platforms like C# and VB.Net with the .Net framework as they seamlessly integrate with the Office suite, with little overhead compared to the direct integration of VBA, and give you access to a wealth of modern technologies.
But don’t be fooled by the FUD about the dangers of keeping a legacy VBA code base and do your due diligence: does the guy who suggest you a full migration will do it for free or is he paid for the job? A full migration may be a necessity: not because the platform is VBA but because the application is buggy, undocumented, out of control and using it creates a true operational risk, and this is true for any application and technology including the newest.
Then, if you have a VBA application that is working perfectly, is documented and controlled, an alternative to both the headlong rush and the full migration is the integration of .Net components with VBA, you then use a future-proof technology to develop new features and/or replace legacy features as you go along, avoiding the big-bang effect.
So now you know what to do and right now I’ll show you how to do this seamless transition from VBA to greener pastures by implementing a simple API with three popular .Net languages: C# (the king), VB.Net (the prince) and C++/CLI (the Jack of all trades, master of none).