jMeter: load multiple rows from CSV file into HTTP body - jmeter

Question

I'm trying to build an HTTP request to a service which accepts a list of items, for example:
{
"user": "john",
"table": "goods",
"articles": [
{"id": "003", "quantity": 1},
{"id": "004", "quantity": 1},
{"id": "023", "quantity": 2},
{"id": "011", "quantity": 3},
{"id": "063", "quantity": 1},
{"id": "006", "quantity": 7}
]
}
My goal is to load ALL the articles from a CSV file, done as follows:
I would like to have something like:
{
"user": "john",
"table": "goods",
"articles": [
{"id": "${id}", "quantity": ${qte}}
]
}
Can anyone help me?
UPDATE:
Dmitri T solved my problem, thank you!
I also asked him how to add a JSON object (a dictionary) to my body, while he answer I found a solution which I will share here in case someone needs it. If you want to add:
"user": {"id": 1, "name": "John"}
to your body content, you just need to do this:
def user = [:]
//populate user
user.put('id', 1)
user.put('name', 'John')
//add user entity to body content
content.put("user", user)
Have fun!

Answer 1

You won't be able to have something like you provided above, the only way to dynamically build a JSON request body is constructing it programmatically using JSR223 PreProcessor
Add JSR223 PreProcessor as a child of the request which body you want to parameterize
Put the following code into "Script" area:
def content = [:]
content.put('user', 'john')
content.put('table', 'goods')
def articles = []
new File('test.csv').readLines().each { line ->
def article = [:]
article.put('id', line.split(',')[0])
article.put('quantity', line.split(',')[1])
articles.add(article)
}
content.put('articles', articles)
sampler.addNonEncodedArgument('', new groovy.json.JsonBuilder(content).toPrettyString(), '')
sampler.setPostBodyRaw(true)
That's it, when you run your test the PreProcessor will generate the request body from the CSV file and add it to the HTTP Request sampler on the fly.
References:
Apache Groovy: Parsing and producing JSON
Apache Groovy - Why and How You Should Use It