Jmeter - reset variables generated from Expression extractor - variables

Question

I'm generating a number of variables in a script via the use of JSON and Regular expression extractors.
I need to reset these for each test record I process to ensure that each one uses the correct extracted value.
The problem (if it is a problem) is that the generated values consist of a number of array elements or groups.
example:
Regular expression extractor generating Reference Name: identifier
This generates:
- Identifier
- Identifier_g
- Identifier_g0
- Identifier_g1
I can successfully reset Identifier using vars.put("Identifier",null); at an appropriate point in a beanshell, however this doesn't reset the other items.
I have a number of these extractions taking place and the number of results they produce can vary, so the 'g' value above may change.
Any advice on how to reset these (or even if I need to)?

Answer 1

If you can reset one variable - you can reset them all, the only thing I would recommend is using Groovy instead of Beanshell. The reasons are in:
Groovy is more Java-compliant, you can use all modern Java features while with Beanshell you have to leave with Java 5 language level
Groovy has a lot of nice JDK enhancements providing some "syntax sugar" to base Java classes
Groovy performance is much better as well-behaved Groovy scripts can be compiled and cached while Beanshell are interpreted each time so Groovy scripts will perform much faster and have lesser memory footprint
References:
JMeter Best Practices: JSR223 Elements
Apache Groovy - Why and How You Should Use It
And finally the relevant Groovy code to reset all the variables whose names start with Identifier
vars.getIterator().each {
def name = it.getKey()
if (name.startsWith('Identifier')) {
vars.put(name, null)
}
}

Answer 2

Whenever the extractor runs again it automatically cleans up old values.
this would only be a problem if you're using same var names in different extractors.

Answer 3

When you use Regular expression and you define Template as $1$ for example it will take only Identifier_g1 value, which means matching group 1 of Regular expression, the other values aren't used so in a normal flow you don't need to care them.
Groups Number are constant, basically the number of round brackets you create in Regular expression (\w+)=(\d+) will create 2 groups so you'll have to delete until Identifier_g2.
If you want you can add JSR 223 element to remove specific values.
You can choose between JSR223 Sampler or JSR223 PreProcessor or JSR223 PostProcessor depending on when you want it to be executed. Notice that Sampler will be included in reports and will marked as fail if script fails.
Code in your case:
vars.remove("Identifier");
vars.remove("Identifier_g");
vars.remove("Identifier_g0");
vars.remove("Identifier_g1");