I'm writing a parser for the output of a legacy application, and since there are no specs on the file syntax I've got as many samples of these files as I could.
Now I'm writing the unit tests before implementing the parser (because there is no other sane way to do this) but I'm not sure whether I should:
use the real files produced by the application, reading from them and comparing the output with the output that I would store in json format in another file.
or create a sample string with the tokens and possibilities I want to test and a dict (this is python) with the expected output.
I'm inclined to use the second alternative because I would test only what I need to, without all the "real-world" data included on the actual files, but I'm afraid I could forget to test for one possibility or another.
What do you think?
My suggestion is to do both. Write a set of integration tests that run through all the files you have with the expected outputs then unit test with your expected inputs to isolate the parsing logic.
I would recommend writing the integration tests first so you write your parser outside in, it might be disparaging to see a bunch of failing tests, but it'll help you isolate your edge cases earlier.
Btw, I think this is a great question. I recently came across something a similar problem which was transforming large xml feeds from an upstream system into a proprietary format. My solution was to write a set of integration black box tests for the full feeds testing things like record counts and other high level success metrics, then break down inputs into smaller and smaller chunks until I was able to test all the permutations of the data. It was only then that I had a good understanding of how to build the parser.
You should be careful using production data in testing scenarios. It could be a disaster if all your users got an email from a test environment, for example. Its also probably unethical in certain scenarios for developers to have access to prod data, even if there is no way for the users to know this. Think medical, bank, college grades types scenarios.
My answer is you should use data that is close to prod data. If you want to use the actual prod data, you need to scrub it for the above scenarios.
Production data can be a good starting point (assuming it's not sensitive info), since there's a good chance you can't think of all the possible permutations yourself. However, once you get a good working set of data, save it somewhere static, like a file. Then have the tests get it from there instead of dynamically from the production environment. That way you can run the tests with a known set of inputs every time.
The alternative, getting production data on the fly for test inputs, is fraught with problems. Changes in the data could cause a test to pass one time, but fail the next because the inputs changed.
Don't forget to structure the test such that you can add additional possibilities (i.e., regression tests) as they become known.
Using the second solution you offer will allow you to control what is expected and what is returned, which is ideal for unit testing. When creating automated tests, it is best to avoid manual interaction as often as possible - visually scanning the results is one of those practices you should avoid when possible (assuming that's what you meant by "compare").
Related
I'm a controls developer and a relative newbie to unit testing. Almost daily, I fight the attitude that you cannot test controls because of the UI interaction. I'm producing a demonstration control to show that it's possible to dramatically reduce manual testing if the control is designed to be testable. Currently I've got 50% logic coverage, but I think I could bump that up to 75% or higher if I could find a way to test some of the more complicated parts.
For example, I have a class with properties that describe the control's state and a method that generates a WPF PathGeometry object made of several segments. The implementation looks something like this:
internal PathGeometry CreateOuterGeometry()
{
double arcRadius = OuterCoordinates.Radius;
double sweepAngle = OuterCoordinates.SweepAngle;
ArcSegment outerArc = new ArcSegment(...);
LineSegment arcEndToCenter = new LineSegment(...);
PathFigure fig = new PathFigure();
// configure figure and add segments...
PathGeometry outerGeometry = new PathGeometry();
outerGeometry.Figures.Add(fig);
return outerGeometry;
}
I've got a few other methods like this that account for a few hundred blocks of uncovered code, an extra 25% coverage. I originally planned to test these methods, but rejected the notion. I'm still a unit testing newbie, and the only way I could think of to test the code would be several methods like this:
void CreateOuterGeometry_AngleIsSmall_ArcSegmentIsCorrect()
{
ClassUnderTest classUnderTest = new ClassUnderTest();
// configure the class under test...
ArcSegment expectedArc = // generate expected Arc...
PathGeometry geometry = classUnderTest.CreateOuterGeometry()
ArcSegment arc = geometry.Figures.Segments[0];
Assert.AreEqual(expectedArc, arc)
}
The test itself looks fine; I'd write one for each expected segment. But I had some problems:
Do I need tests to verify "Is the first segment an ArcSegment?" In theory the test tests this, but shouldn't each test only test one thing? This sounds like two things.
The control has at least six cases for calculation and four edge cases; this means for each method I need at least ten tests.
During development I changed how the various geometries were generated several times. This would cause me to have to rewrite all of the tests.
The first problem gave me pause because it seemed like it might inflate the number of tests. I thought I might have to test things like "Were there x segments?" and "Is segment n the right type?", but now that I've thought more I see that there's no branching logic in the method so I only need to do those tests once. The second problem made me more confident that there would be much effort associated with the test. It seems unavoidable. The third problem compounds the first two. Every time I changed the way the geometry was calculated, I'd have to edit an estimated 40 tests to make them respect the new logic. This would also include adding or removing tests if segments were added or removed.
Because of these three problems, I opted to write an application and manual test plan that puts the control in all of the interesting states and asks the user to verify it looks a particular way. Was this wrong? Am I overestimating the effort involved with writing the unit tests? Is there an alternative way to test this that might be easier? (I'm currently studying mocks and stubs; it seems like it'd require some refactoring of the design and end up being approximately as much effort.)
Use dependency injection and mocks.
Create interfaces for ArcSegmentFactory, LineSegmentFactory, etc., and pass a mock factory to your class. This way, you'll isolate the logic that is specific to this object (this should make testing easier), and won't be depending on the logic of your other objects.
About what to test: you should test what's important. You probably have a timeline in which you want to have things done, and you probably won't be able to test every single thing. Prioritize stuff you need to test, and test in order of priority (considering how much time it will take to test). Also, when you've already made some tests, it gets much easier to create new tests for other stuff, and I don't really see a problem in creating multiple tests for the same class...
About the changes, that's what tests are for: allowing you to change and don't really fear your change will bring chaos to the world.
You might try writing a control generation tool that generates random control graphs, and test those. This might yield some data points that you might not have thought of.
In our project, we use JUnit to perform tests which are not, strictly speaking, unit tests. We find, for example, that it's helpful to hook up a blank database and compare an automatic schema generated by Hibernate (an Object-Relational Mapping tool) to the actual schema for our test database; this helps us catch a lot of issues with wrong database mappings. But in general... you should only be testing one method, on one class, in a given test method. That doesn't mean you can't do multiple assertions against it to examine various properties of the object.
My approach is to convert the graph into a string (one segment per line) and compare this string to an expected result.
If you change something in your code, tests will start to fail but all you need to do is to check that the failures are in the right places. Your IDE should offer a side-by-side diff for this.
When you're confident that the new output is correct, just copy it over the old expected result. This will make sure that a mistake won't go unnoticed (at least not for long), the tests will still be simple and they are quick to fix.
Next, if you have common path parts, then you can put them into individual strings and build the expected result of a test from those parts. This allows you to avoid repeating yourself (and if the common part changes, you just have to update a single place for all tests).
If I understand your example correctly, you were trying to find a way to test whether a whole bunch of draw operations produce a given result.
Instead of human eyes, you could have produced a set of expected images (a snapshot of verified "good" images), and created unit tests which use the draw operations to create the same set of images and compare the result with an image comparison. This would allow you to automate the testing of the graphic operations, which is what I understand your problem to be.
The textbook way to do this would be to move all the business logic to libraries or controllers which are called by a 1 line method in the GUI. That way you can unit test the controller or library without dealing with the GUI.
This question is more or less programming language agnostic. However as I'm mostly into Java these days that's where I'll draw my examples from. I'm also thinking about the OOP case, so if you want to test a method you need an instance of that methods class.
A core rule for unit tests is that they should be autonomous, and that can be achieved by isolating a class from its dependencies. There are several ways to do it and it depends on if you inject your dependencies using IoC (in the Java world we have Spring, EJB3 and other frameworks/platforms which provide injection capabilities) and/or if you mock objects (for Java you have JMock and EasyMock) to separate a class being tested from its dependencies.
If we need to test groups of methods in different classes* and see that they are well integration, we write integration tests. And here is my question!
At least in web applications, state is often persisted to a database. We could use the same tools as for unit tests to achieve independence from the database. But in my humble opinion I think that there are cases when not using a database for integration tests is mocking too much (but feel free to disagree; not using a database at all, ever, is also a valid answer as it makes the question irrelevant).
When you use a database for integration tests, how do you fill that database with data? I can see two approaches:
Store the database contents for the integration test and load it before starting the test. If it's stored as an SQL dump, a database file, XML or something else would be interesting to know.
Create the necessary database structures by API calls. These calls are probably split up into several methods in your test code and each of these methods may fail. It could be seen as your integration test having dependencies on other tests.
How are you making certain that database data needed for tests is there when you need it? And why did you choose the method you choose?
Please provide an answer with a motivation, as it's in the motivation the interesting part lies. Remember that just saying "It's best practice!" isn't a real motivation, it's just re-iterating something you've read or heard from someone. For that case please explain why it's best practice.
*I'm including one method calling other methods in (the same or other) instances of the same class in my definition of unit test, even though it might technically not be correct. Feel free to correct me, but let's keep it as a side issue.
I prefer creating the test data using API calls.
In the beginning of the test, you create an empty database (in-memory or the same that is used in production), run the install script to initialize it, and then create whatever test data used by the database. Creation of the test data may be organized for example with the Object Mother pattern, so that the same data can be reused in many tests, possibly with minor variations.
You want to have the database in a known state before every test, in order to have reproducable tests without side effects. So when a test ends, you should drop the test database or roll back the transaction, so that the next test could recreate the test data always the same way, regardless of whether the previous tests passed or failed.
The reason why I would avoid importing database dumps (or similar), is that it will couple the test data with the database schema. When the database schema changes, you would also need to change or recreate the test data, which may require manual work.
If the test data is specified in code, you will have the power of your IDE's refactoring tools at your hand. When you make a change which affects the database schema, it will probably also affect the API calls, so you will anyways need to refactor the code using the API. With nearly the same effort you can also refactor the creation of the test data - especially if the refactoring can be automated (renames, introducing parameters etc.). But if the tests rely on a database dump, you would need to manually refactor the database dump in addition to refactoring the code which uses the API.
Another thing related to integration testing the database, is testing that upgrading from a previous database schema works right. For that you might want to read the book Refactoring Databases: Evolutionary Database Design or this article: http://martinfowler.com/articles/evodb.html
In integration tests, you need to test with real database, as you have to verify that your application can actually talk to the database. Isolating the database as dependency means that you are postponing the real test of whether your database was deployed properly, your schema is as expected and your app is configured with the right connection string. You don't want to find any problems with these when you deploy to production.
You also want to test with both precreated data sets and empty data set. You need to test both path where your app starts with an empty database with only your default initial data and starts creating and populating the data and also with a well-defined data sets that target specific conditions you want to test, like stress, performance and so on.
Also, make sur that you have the database in a well-known state before each state. You don't want to have dependencies between your integration tests.
Why are these two approaches defined as being exclusively?
I can't see any viable argument for
not using pre-existing data sets, especially particular data that has
caused problems in the past.
I can't
see any viable argument for not
programmatically extending that data with
all the possible conditions that
you can imagine causing problems and even a
bit of random data for integration
testing.
In modern agile approaches, Unit tests are where it really matters that the same tests are run each time. This is because unit tests are aimed not at finding bugs but at preserving the functionality of the app as it is developed, allowing the developer to refactor as needed.
Integration tests, on the other hand, are designed to find the bugs you did not expect. Running with some different data each time can even be good, in my opinion. You just have to make sure your test preserves the failing data if you get a failure. Remember, in formal integration testing, the application itself will be frozen except for bug fixes so your tests can be change to test for the maximum possible number and kinds of bugs. In integration, you can and should throw the kitchen sink at the app.
As others have noted, of course, all this naturally depends on the kind of application that you are developing and the kind of organization you are in, etc.
It sounds like your question is actually two questions. Should you exclude the database from your testing? When you do a database, then how should you generate the data in the database?
When possible I prefer to use an actual database. Frequently the queries (written in SQL, HQL, etc.) in CRUD classes can return surprising results when confronted with an actual database. It's better to flush these issues out early on. Often developers will write very thin unit tests for CRUD; testing only the most benign cases. Using an actual database for your testing can test all kinds corner cases you may not have even been aware of.
That being said there can be other issues. After each test you want to return your database to a known state. It my current job we nuke the database by executing all the DROP statements and then completely recreating all the tables from scratch. This is extremely slow on Oracle, but can be very fast if you use an in memory database like HSQLDB. When we need to flush out Oracle specific issues we just change the database URL and driver properties and then run against Oracle. If you don't have this kind of database portability then Oracle also has some kind of database snapshot feature which can be used specifically for this purpose; rolling back the entire database to some previous state. I'm sure what other databases have.
Depending on what kind of data will be in your database the API or the load approach may work better or worse. When you have highly structured data with many relations, APIs will make your life easier my making the relations between your data explicit. It will be harder for you to make a mistake when creating your test data set. As mentioned by other posters refactoring tools can take care of some of the changes to structure of your data automatically. Often I find it useful to think of API generated test data as composing a scenario; when a user/system has done steps X, Y Z and then tests will go from there. These states can be achieved because you can write a program that calls the same API your user would use.
Loading data becomes much more important when you need large volumes of data, you have few relations between within your data or there is consistency in the data that can not be expressed using APIs or standard relational mechanisms. At one job that at worked at my team was writing the reporting application for a large network packet inspection system. The volume of data was quite large for the time. In order to trigger a useful subset of test cases we really needed test data generated by the sniffers. This way correlations between the information about one protocol would correlate with information about another protocol. It was difficult to capture this in the API.
Most databases have tools to import and export delimited text files of tables. But often you only want subsets of them; making using data dumps more complicated. At my current job we need to take some dumps of actual data which gets generated by Matlab programs and stored in the database. We have tool which can dump a subset of the database data and then compare it with the "ground truth" for testing. It seems our extraction tools are being constantly modified.
I've used DBUnit to take snapshots of records in a database and store them in XML format. Then my unit tests (we called them integration tests when they required a database), can wipe and restore from the XML file at the start of each test.
I'm undecided whether this is worth the effort. One problem is dependencies on other tables. We left static reference tables alone, and built some tools to detect and extract all child tables along with the requested records. I read someone's recommendation to disable all foreign keys in your integration test database. That would make it way easier to prepare the data, but you're no longer checking for any referential integrity problems in your tests.
Another problem is database schema changes. We wrote some tools that would add default values for columns that had been added since the snapshots were taken.
Obviously these tests were way slower than pure unit tests.
When you're trying to test some legacy code where it's very difficult to write unit tests for individual classes, this approach may be worth the effort.
I do both, depending on what I need to test:
I import static test data from SQL scripts or DB dumps. This data is used in object load (deserialization or object mapping) and in SQL query tests (when I want to know whether the code will return the correct result).
Plus, I usually have some backbone data (config, value to name lookup tables, etc). These are also loaded in this step. Note that this loading is a single test (along with creating the DB from scratch).
When I have code which modifies the DB (object -> DB), I usually run it against a living DB (in memory or a test instance somewhere). This is to ensure that the code works; not to create any large amount of rows. After the test, I rollback the transaction (following the rule that tests must not modify the global state).
Of course, there are exceptions to the rule:
I also create large amount of rows in performance tests.
Sometimes, I have to commit the result of a unit test (otherwise, the test would grow too big).
I generally use SQL scripts to fill the data in the scenario you discuss.
It's straight-forward and very easily repeatable.
This will probably not answer all your questions, if any, but I made the decision in one project to do unit testing against the DB. I felt in my case that the DB structure needed testing too, i.e. did my DB design deliver what is needed for the application. Later in the project when I feel the DB structure is stable, I will probably move away from this.
To generate data I decided to create an external application that filled the DB with "random" data, I created a person-name and company-name generators etc.
The reason for doing this in an external program was:
1. I could rerun the tests on by test modified data, i.e. making sure my tests were able to run several times and the data modification made by the tests were valid modifications.
2. I could if needed, clean the DB and get a fresh start.
I agree that there are points of failure in this approach, but in my case since e.g. person generation was part of the business logic generating data for tests was actually testing that part too.
Our team confront the same question recently.
Before, we were using specflow to do integration testing. With specflow, QA can write each test case inside which populating necessary test data to DB.
Now, QA want to use postman to test API, how can they populate the data? One solution is creating Apis for populating them. Another is sync historical data from Prod to test env.
Will update my answer once we try different solutions and decide which one to go.
How would you "unit test" a report created by some report engine like Crystal Reports or SQL Server Reporting Services?
The problem with reports is akin to the problem with GUIs.
If the Report/GUI has lot of (misplaced) intelligence it is going to make testing difficult.
The solution then is to
Separated Presentation : Separate presentation from content (data-access/domain/business rules). In the current context would mean, that you create some sort of ViewModel class that mirrors the content of the final report (e.g. if you have order details and line items in your report, this class should have properties for the details and a list of line item objects). The ViewModel is infinitely simpler to test. The last-mile, applying presentation to the content should be relatively trivial (thin UI).
e.g. if you use xslt to render your reports, you can test the data xml using tools like XmlUnit or string compare. You can reasonable confident in xsl transformations on the data xml for the final report... Also any bugs in here would be trivial to fix.
However if you're using third party vendors like Crystal Reports, you have no control / access to hook in to the report generation. In such cases, the best you can do is generate representative/expected output files (e.g. pdfs) called Golden Files. Use this as a read-only resource in your tests to compare the actual output. However this approach is very fragile.. in that any substantial change to the report generation code might render all previous Golden Files incorrect. So they would have to be regenerated. If the cost to benefit ratio of automation is too high, I'd say Go manual with old-school word doc test plans.
For testing our own Java-based reporting product, i-net Clear Reports, we run a whole slew of test reports once, exporting them to various export formats, make sure the output is as desired, and then continously have these same reports run daily, comparing the results to the original data. Any differences then show up as test failures.
It has worked pretty well for us. Disadvantage of this is any minor differences that might not make any difference show up as test failures until the test data is reset.
Side note: this isn't exactly a unit test but rather an acceptance test. But I don't see how you could truly "unit test" an entire report.
The best I can think of, is comparing the results to an expected output.
Maybe some intelligence can be added, but it is not that easy to test these big blocks.
I agree with Gamecat.
Generate the report from fixed (constant) data, and compare it to the expected output for that data.
After that you might be able to use simple tests such as diff (checking if the files are identical)
My current idea is to create tests at two levels:
Unit tests: Structure the report to enable testing using some ideas for testing a UI, like Humble View. The report itself will be made as dumb as possible. It should consist mostly of just simple field bindings. The data items/objects that act as the source of these bindings can then be unit tested.
Acceptence tests: Generate some example reports. Verify them by hand first. Then setup an automated test that does a compare using diff.
Ok, maybe I'm missing something, but I really don't see the point of Selenium. What is the point of opening the browser using code, clicking buttons using code, and checking for text using code? I read the website and I see how in theory it would be good to automatically unit test your web applications, but in the end doesn't it just take much more time to write all this code rather than just clicking around and visually verifying things work?
I don't get it...
It allows you to write functional tests in your "unit" testing framework (the issue is the naming of the later).
When you are testing your application through the browser you are usually testing the system fully integrated. Consider you already have to test your changes before committing them (smoke tests), you don't want to test it manually over and over.
Something really nice, is that you can automate your smoke tests, and QA can augment those. Pretty effective, as it reduces duplication of efforts and gets the whole team closer.
Ps as any practice that you are using the first time it has a learning curve, so it usually takes longer the first times. I also suggest you look at the Page Object pattern, it helps on keeping the tests clean.
Update 1: Notice that the tests will also run javascript on the pages, which helps testing highly dynamic pages. Also note that you can run it with different browsers, so you can check cross-browser issues(at least on the functional side, as you still need to check the visual).
Also note that as the amount of pages covered by tests builds up, you can create tests with complete cycles of interactions quickly. Using the Page Object pattern they look like:
LastPage aPage = somePage
.SomeAction()
.AnotherActionWithParams("somevalue")
//... other actions
.AnotherOneThatKeepsYouOnthePage();
// add some asserts using methods that give you info
// on LastPage (or that check the info is there).
// you can of course break the statements to add additional
// asserts on the multi-steps story.
It is important to understand that you go gradual about this. If it is an already built system, you add tests for features/changes you are working on. Adding more and more coverage along the way. Going manual instead, usually hides what you missed to test, so if you made a change that affects every single page and you will check a subset (as time doesn't allows), you know which ones you actually tested and QA can work from there (hopefully by adding even more tests).
This is a common thing that is said about unit testing in general. "I need to write twice as much code for testing?" The same principles apply here. The payoff is the ability to change your code and know that you aren't breaking anything.
Because you can repeat the SAME test over and over again.
If your application is even 50+ pages and you need to do frequent builds and test it against X number of major browsers it makes a lot of sense.
Imagine you have 50 pages, all with 10 links each, and some with multi-stage forms that require you to go through the forms, putting in about 100 different sets of information to verify that they work properly with all credit card numbers, all addresses in all countries, etc.
That's virtually impossible to test manually. It becomes so prone to human error that you can't guarantee the testing was done right, never mind what the testing proved about the thing being tested.
Moreover, if you follow a modern development model, with many developers all working on the same site in a disconnected, distributed fashion (some working on the site from their laptop while on a plane, for instance), then the human testers won't even be able to access it, much less have the patience to re-test every time a single developer tries something new.
On any decent size of website, tests HAVE to be automated.
The point is the same as for any kind of automated testing: writing the code may take more time than "just clicking around and visually verifying things work", maybe 10 or even 50 times more.
But any nontrivial application will have to be tested far more than 50 times eventually, and manual tests are an annoying chore that will likely be omitted or done shoddily under pressure, which results in bugs remaining undiscovered until just bfore (or after) important deadlines, which results in stressful all-night coding sessions or even outright monetary loss due to contract penalties.
Selenium (along with similar tools, like Watir) lets you run tests against the user interface of your Web app in ways that computers are good at: thousands of times overnight, or within seconds after every source checkin. (Note that there are plenty of other UI testing pieces that humans are much better at, such as noticing that some odd thing not directly related to the test is amiss.)
There are other ways to involve the whole stack of your app by looking at the generated HTML rather than launching a browser to render it, such as Webrat and Mechanize. Most of these don't have a way to interact with JavaScript-heavy UIs; Selenium has you somewhat covered here.
Selenium will record and re-run all of the manual clicking and typing you do to test your web application. Over and over.
Over time studies of myself have shown me that I tend to do fewer tests and start skipping some, or forgetting about them.
Selenium will instead take each test, run it, if it doesn't return what you expect it, it can let you know.
There is an upfront cost of time to record all these tests. I would recommend it like unit tests -- if you don't have it already, start using it with the most complex, touchy, or most updated parts of your code.
And if you save those tests as JUnit classes you can rerun them at your leisure, as part of your automated build, or in a poor man's load test using JMeter.
In a past job we used to unit test our web-app. If the web-app changes its look the tests don't need to be re-written. Record-and-replay type tests would all need to be re-done.
Why do you need Selenium? Because testers are human beings. They go home every day, can't always work weekends, take sickies, take public holidays, go on vacation every now and then, get bored doing repetitive tasks and can't always rely on them being around when you need them.
I'm not saying you should get rid of testers, but an automated UI testing tool complements system testers.
The point is the ability to automate what was before a manual and time consuming test. Yes, it takes time to write the tests, but once written, they can be run as often as the team wishes. Each time they are run, they are verifying that behavior of the web application is consistent. Selenium is not a perfect product, but it is very good at automating realistic user interaction with a browser.
If you do not like the Selenium approach, you can try HtmlUnit, I find it more useful and easy to integrate into existing unit tests.
For applications with rich web interfaces (like many GWT projects) Selenium/Windmill/WebDriver/etc is the way to create acceptance tests. In case of GWT/GXT, the final user interface code is in JavaScript so creating acceptance tests using normal junit test cases is basically out of question. With Selenium you can create test scenarios matching real user actions and expected results.
Based on my experience with Selenium it can reveal bugs in the application logic and user interface (in case your test cases are well written). Dealing with AJAX front ends requires some extra effort but it is still feasible.
I use it to test multi page forms as this takes the burden out of typing the same thing over and over again. And having the ability to check if certain elements are present is great. Again, using the form as an example your final selenium test could check if something like say "Thanks Mr. Rogers for ordering..." appears at the end of the ordering process.
How are people unit testing their business applications? I've seen a lot of examples of unit testing with "simple to test" examples. Ex. a calculator. How are people unit testing data-heavy applications? How are you putting together your sample data? In many cases, data for one test may not work at all for another test which makes it hard to just have one test database?
Testing the data access portion of the code is fairly straightforward. It's testing out all the methods that work against the data that seem to be hard to test. For example, imagine a posting process where there is heavy data access to determine what is posted, numbers are adjusted, etc. There are a number of interim steps that occur (and need to be tested) along with tests afterwards that ensure the posting was successful. Some of those steps may actually be stored procedures.
In the past I've tried inserting the test data in a test database, then running the test, but honestly it's pretty painful to write this kind of code (and error prone). I've also tried just building a test database up front and rolling back the changes. That works OK but in a number of places you can't easily do this either (and many people would say that's integration testing; so be it, I still need to be able to test this somehow).
If the answer is that there isn't a nice way of handling this and it currently just sort of sucks, that would be useful to know as well.
Any thoughts, ideas, suggestions, or tips are appreciated.
My automated functional tests usually follow one of two patters:
Database Connected Tests
Mock Persistence Layer Tests
Database Connected Tests
When I have automated tests that are connected to the database, I usually make a single test database template that has enough data for all the tests. When the automated tests are run, a new test database is generated from the template for every test. The test database has to be constantly re-generated because test will often change the data. As tests are added, I usually append more data to the test database template.
There are some nice advantages to this testing method. The obvious advantage is that the tests also exercise your schema. Another advantage is that after setting up the initial tests, most new tests will be able to re-use the existing test data. This makes it easy to add more tests.
The downside is that the test database will become unwieldy. Because data will usually be added one test at time, it will be inconsistent and maybe even unrealistic. You will also end up cursing the person who setup the test database when there is a significant database schema change (which for me usually means I end up cursing myself).
This style of testing obviously doesn't work if you can't generate new test databases at will.
Mock Persistence Layer Tests
For this pattern, you create mock objects that live with the test cases. These mock objects intercept the calls to the database so that you can programmatically provide the appropriate results. Basically, when the code you're testing calls the findCustomerByName() method, your mock object is called instead of the persistence layer.
The nice thing about using mock object tests is that you can get very specific. Often times, there are execution paths that you simply can't reach in automated tests w/o mock objects. They also free you from maintaining a large, monolithic set of test data.
Another benefit is the lack of external dependencies. Because the mock objects simulate the persistence layer, your tests are no longer dependent on the database. This is often the deciding factor when choosing which pattern to choose. Mock objects seem to get more traction when dealing with legacy database systems or databases with stringent licensing terms.
The downside of mock objects is that they often result in a lot of extra test code. This isn't horrible because almost any amount of testing code is cheap when amortized over the number of times you run the test, but it can be annoying to have more test code then production code.
I have to second the comment by #Phil Bennett as I try to approach these integration tests with a rollback solution.
I have a very detailed post about integration testing your data access layer here
I show not only the sample data access class, base class, and sample DB transaction fixture class, but a full CRUD integration test w/ sample data shown. With this approach you don't need multiple test databases as you can control the data going in with each test and after the test is complete the transactions are all rolledback so your DB is clean.
About unit testing business logic inside your app, I would also second the comments by #Phil and #Mark because if you mock out all the dependencies your business object has, it becomes very simple to test your application logic one entity at a time ;)
Edit: So are you looking for one huge integration test that will verify everything from logic pre-data base / stored procedure run w/ logic and finally a verification on the way back? If so you could break this out into 2 steps:
1 - Unit test the logic that happens before the data is pushed
into your data access code. For
example, if you have some code that
calculates some numbers based on
some properties -- write a test that
only checks to see if the logic for
this 1 function does what you asked
it to do. Mock out any dependancy
on the data access class so you can
ignore it for this test of the
application logic alone.
2 - Integration test the logic that happens once you take your
manipulated data (from the previous
method we unit tested) and call the
appropriate stored procedure. Do
this inside a data specific testing
class so you can rollback after it's
completed. After your stored
procedure has run, do a query
against the database to get your
object now that we have done some
logic against the data and verify it
has the values you expected
(post-stored procedure logic /etc )
If you need an entry in your database for the stored procedure to run, simply insert that data before you run the sproc that has your logic inside it. For example, if you have a product that you need to test, it might require a supplier and category entry to insert so before you insert your product do a quick and dirty insert for a supplier and category so your product insert works as planned.
It depends on what you're testing. If you're testing a business logic component -- then its immaterial where the data is coming from and you'd probably use a mock or a hand rolled stub class that simulates the data access routine the component would have called in the wild. The only time I mess with the data access is when I'm actually testing the data access components themselves.
Even then I tend to open a DB transaction in the TestFixtureSetUp method (obviously this depends on what unit testing framework you might be using) and rollback the transaction at the end of the test suite TestFixtureTeardown.
Mocking Frameworks enable you to test your business objects.
Data Driven tests often end up becoming more of a intergration test than a unit test, they also carry with them the burden of managing the state of a data store pre and post execution of the test and the time taken in connecting and executing queries.
In general i would avoid doing unit tests that touch the database from your business objects. As for Testing your database you need a different stratergy.
That being said you can never totally get away from data driven testing only limiting the amout of tests that actually need to invoke your back end systems.
It sounds like you might be testing message based systems, or systems with highly parameterised interfaces, where there are large numbers of permutations of input data.
In general all the rules of standard unti testing still hold:
Try to make the units being tested as small and discrete as possible.
Try to make tests independant.
Factor code to decouple dependencies.
Use mocks and stubs to replace dependencies (like dataaccess)
Once this is done you will have removed a lot of the complexity from the tests, hopefully revealing good sets of unit tests, and simplifying the sample data.
A good methodology for then compiling sample data for test that still require complex input data is Orthogonal testing, or see here.
I've used that sort of method for generating test plans for WCF and BizTalk solutions where the permutations of input messages can create multiple possible execution paths.
For lots of different runs over the same logic but with different data you can use CSV, as many columns as you like for the input and the last for the output etc.