Parameterization & Data-Driven Tests

Hardcoded test data is a trap. When you test checkout with the same user and same product 1,000 times, you're not simulating reality. Real e-commerce has thousands of users buying thousands of products.

Parameterization separates your test logic from test data, enabling realistic, data-driven performance tests.

The Problem with Hardcoded Data

<!-- Bad: Same user, same product, every iteration -->
<HTTPSamplerProxy testname="Login">
  <stringProp name="Argument.value">{"email":"test@example.com","password":"test123"}</stringProp>
</HTTPSamplerProxy>

<HTTPSamplerProxy testname="Add to Cart">
  <stringProp name="HTTPSampler.path">/api/cart/add/PROD-001</stringProp>
</HTTPSamplerProxy>

Problems:

Server may cache results for repeated identical requests
Same user session conflicts across threads
Database hotspots on same product row
Doesn't test inventory logic with different products

CSV Data Set Config

The most powerful parameterization tool in JMeter. Reads data from external CSV files.

Create Your Data File

users.csv:

email,password,firstName
john@example.com,john123,John
jane@example.com,jane456,Jane
mike@example.com,mike789,Mike
sara@example.com,sara012,Sara

products.csv:

productId,productName,expectedPrice
PROD-001,Wireless Headphones,79.99
PROD-002,USB-C Hub,49.99
PROD-003,Mechanical Keyboard,129.99
PROD-004,4K Monitor,399.99
PROD-005,Laptop Stand,34.99

Configure CSV Data Set

<CSVDataSet guiclass="TestBeanGUI" testclass="CSVDataSet"
            testname="User Data">
  <stringProp name="filename">users.csv</stringProp>
  <stringProp name="fileEncoding">UTF-8</stringProp>
  <stringProp name="variableNames">email,password,firstName</stringProp>
  <stringProp name="delimiter">,</stringProp>
  <boolProp name="ignoreFirstLine">true</boolProp>
  <boolProp name="quotedData">false</boolProp>
  <boolProp name="recycle">true</boolProp>
  <boolProp name="stopThread">false</boolProp>
  <stringProp name="shareMode">shareMode.all</stringProp>
</CSVDataSet>

Property	Value	Purpose
Filename	`users.csv`	Path to CSV file
Variable Names	`email,password,firstName`	Column-to-variable mapping
Ignore First Line	`true`	Skip header row
Recycle on EOF	`true/false`	Restart from beginning when file ends
Stop Thread on EOF	`false`	What to do when data runs out
Sharing Mode	`All threads`	How threads share data

Use Variables in Requests

<HTTPSamplerProxy testname="Login">
  <stringProp name="HTTPSampler.method">POST</stringProp>
  <stringProp name="HTTPSampler.path">/api/auth/login</stringProp>
  <boolProp name="HTTPSampler.postBodyRaw">true</boolProp>
  <stringProp name="Argument.value">{
    "email": "${email}",
    "password": "${password}"
  }</stringProp>
</HTTPSamplerProxy>

<HTTPSamplerProxy testname="Add Product to Cart">
  <stringProp name="HTTPSampler.method">POST</stringProp>
  <stringProp name="HTTPSampler.path">/api/cart/items</stringProp>
  <boolProp name="HTTPSampler.postBodyRaw">true</boolProp>
  <stringProp name="Argument.value">{
    "productId": "${productId}",
    "quantity": 1
  }</stringProp>
</HTTPSamplerProxy>

Sharing Modes Explained

All Threads (Default)

All threads share one file cursor. Each thread gets the next row.

Thread 1: Gets row 1 (john@example.com)
Thread 2: Gets row 2 (jane@example.com)
Thread 3: Gets row 3 (mike@example.com)
Thread 1 (iteration 2): Gets row 4 (sara@example.com)

Use for: Unique user credentials, order numbers, anything that shouldn't repeat.

Current Thread Group

Each thread group has its own cursor. Threads within a group share.

Thread Group 1:
  Thread 1: Gets row 1
  Thread 2: Gets row 2

Thread Group 2:
  Thread 1: Gets row 1 (same as TG1-T1!)
  Thread 2: Gets row 2 (same as TG1-T2!)

Use for: When different thread groups need independent data sets.

Current Thread

Each thread has its own cursor. Every thread reads the entire file independently.

Thread 1: Gets row 1, then row 2, then row 3...
Thread 2: Gets row 1, then row 2, then row 3...

Use for: When you want every thread to cycle through all test data.

User Defined Variables

For configuration values that don't change during test execution.

<Arguments guiclass="ArgumentsPanel" testclass="Arguments"
           testname="User Defined Variables">
  <collectionProp name="Arguments.arguments">
    <elementProp name="BASE_URL" elementType="Argument">
      <stringProp name="Argument.name">BASE_URL</stringProp>
      <stringProp name="Argument.value">https://api.mystore.com</stringProp>
    </elementProp>
    <elementProp name="API_VERSION" elementType="Argument">
      <stringProp name="Argument.name">API_VERSION</stringProp>
      <stringProp name="Argument.value">v2</stringProp>
    </elementProp>
    <elementProp name="TIMEOUT" elementType="Argument">
      <stringProp name="Argument.name">TIMEOUT</stringProp>
      <stringProp name="Argument.value">30000</stringProp>
    </elementProp>
  </collectionProp>
</Arguments>

Use in requests:

<HTTPSamplerProxy testname="Get Products">
  <stringProp name="HTTPSampler.domain">${BASE_URL}</stringProp>
  <stringProp name="HTTPSampler.path">/api/${API_VERSION}/products</stringProp>
</HTTPSamplerProxy>

Environment-Specific Config

Pass variables from command line for different environments:

# Staging
jmeter -n -t test.jmx -JBASE_URL=https://staging-api.mystore.com

# Production
jmeter -n -t test.jmx -JBASE_URL=https://api.mystore.com

Access with ${__P(BASE_URL)} function.

Complete E-commerce Data Setup

File Structure

test-data/
├── users.csv           (1000 unique users)
├── products.csv        (500 products)
├── addresses.csv       (user shipping addresses)
├── payment_methods.csv (test credit cards)
└── promo_codes.csv     (discount codes)

users.csv

email,password,firstName,lastName,userId
user001@test.com,Pass001!,John,Smith,USR001
user002@test.com,Pass002!,Jane,Doe,USR002
user003@test.com,Pass003!,Bob,Johnson,USR003

products.csv

productId,sku,category,price,stockLevel
PROD-001,SKU-WH-001,Electronics,79.99,500
PROD-002,SKU-KB-002,Electronics,129.99,200
PROD-003,SKU-SH-003,Footwear,89.99,1000

Test Plan Configuration

<!-- User data - unique per thread, don't recycle -->
<CSVDataSet testname="User Data">
  <stringProp name="filename">test-data/users.csv</stringProp>
  <stringProp name="variableNames">email,password,firstName,lastName,userId</stringProp>
  <boolProp name="ignoreFirstLine">true</boolProp>
  <boolProp name="recycle">false</boolProp>
  <boolProp name="stopThread">true</boolProp>
  <stringProp name="shareMode">shareMode.all</stringProp>
</CSVDataSet>

<!-- Product data - can recycle, shared across threads -->
<CSVDataSet testname="Product Data">
  <stringProp name="filename">test-data/products.csv</stringProp>
  <stringProp name="variableNames">productId,sku,category,price,stockLevel</stringProp>
  <boolProp name="ignoreFirstLine">true</boolProp>
  <boolProp name="recycle">true</boolProp>
  <boolProp name="stopThread">false</boolProp>
  <stringProp name="shareMode">shareMode.all</stringProp>
</CSVDataSet>

Random Selection Functions

Random Number

<!-- Random quantity between 1 and 5 -->
<stringProp name="Argument.value">{
  "productId": "${productId}",
  "quantity": ${__Random(1,5)}
}</stringProp>

Random from List

// JSR223 PreProcessor
def categories = ["Electronics", "Clothing", "Home", "Sports"]
def randomCategory = categories[new Random().nextInt(categories.size())]
vars.put("randomCategory", randomCategory)

Random Product from CSV

Use __CSVRead function for random access:

<!-- Read random row from products.csv -->
<stringProp name="HTTPSampler.path">/api/products/${__CSVRead(products.csv,0)}</stringProp>

Data Generation with Groovy

For dynamic data that doesn't come from files:

// JSR223 PreProcessor - Generate realistic order data
import java.time.LocalDateTime
import java.time.format.DateTimeFormatter

// Random order reference
def orderRef = "ORD-" + System.currentTimeMillis() + "-" + Thread.currentThread().getId()
vars.put("orderRef", orderRef)

// Future delivery date (3-7 days from now)
def deliveryDays = new Random().nextInt(5) + 3
def deliveryDate = LocalDateTime.now().plusDays(deliveryDays)
vars.put("deliveryDate", deliveryDate.format(DateTimeFormatter.ISO_DATE))

// Random shipping option
def shippingOptions = ["standard", "express", "overnight"]
def shipping = shippingOptions[new Random().nextInt(shippingOptions.size())]
vars.put("shippingOption", shipping)

log.info("Generated order: ${orderRef}, delivery: ${vars.get('deliveryDate')}")

Parallel-Safe Data Strategies

Problem: Data Collisions

100 threads using 50 products = guaranteed collisions. Two threads might:

Add same product to cart simultaneously
Both try to buy last item in stock
Cause race conditions in inventory

Solution 1: Unique Data Per Thread

Ensure data file has at least threads * iterations unique rows:

Threads: 100
Iterations: 10
Required rows: 1000 unique users

<CSVDataSet testname="Unique Users">
  <boolProp name="recycle">false</boolProp>
  <boolProp name="stopThread">true</boolProp>
  <stringProp name="shareMode">shareMode.all</stringProp>
</CSVDataSet>

Solution 2: Thread-Specific Data Files

Split data into per-thread files:

users_thread1.csv
users_thread2.csv
...
users_thread100.csv

// JSR223 PreProcessor
def threadNum = ctx.getThreadNum()
def filename = "test-data/users_thread${threadNum}.csv"
vars.put("userDataFile", filename)

Solution 3: Counter for Uniqueness

Use JMeter's Counter element:

<CounterConfig guiclass="CounterConfigGui" testclass="CounterConfig"
               testname="Order Counter">
  <stringProp name="CounterConfig.start">1</stringProp>
  <stringProp name="CounterConfig.end">999999</stringProp>
  <stringProp name="CounterConfig.incr">1</stringProp>
  <stringProp name="CounterConfig.name">orderNumber</stringProp>
  <stringProp name="CounterConfig.format">ORD-000000</stringProp>
  <boolProp name="CounterConfig.per_user">false</boolProp>
</CounterConfig>

per_user=false means all threads share one counter (guaranteed unique).

Key Takeaways

CSV Data Set Config is essential for realistic tests with varied data.
Sharing mode matters: Use "All threads" for unique data, "Current thread" when each thread needs full data set.
Plan your data: Ensure enough unique rows for threads * iterations.
Generate dynamic data: Use Groovy for order numbers, timestamps, and random selections.

Next Lesson

Your tests now use realistic, varied data. But how do you know if responses are correct? A 200 status code doesn't mean success if the response body contains an error message. Next: Assertions - validating that your application actually works under load.

Parameterization & Data-Driven Tests

The Problem with Hardcoded Data

CSV Data Set Config

Create Your Data File

Configure CSV Data Set

Use Variables in Requests

Sharing Modes Explained

All Threads (Default)

Current Thread Group

Current Thread

User Defined Variables

Environment-Specific Config

Complete E-commerce Data Setup

File Structure

users.csv

products.csv

Test Plan Configuration

Random Selection Functions

Random Number

Random from List

Random Product from CSV

Data Generation with Groovy

Parallel-Safe Data Strategies

Problem: Data Collisions

Solution 1: Unique Data Per Thread

Solution 2: Thread-Specific Data Files

Solution 3: Counter for Uniqueness

Key Takeaways

Next Lesson

🧠 Quick Quiz

What is the purpose of CSV Data Set Config?

What does 'Sharing mode: All threads' mean?

Why should you use 'Recycle on EOF: False' for user credentials?

Launch Offer

Parameterization & Data-Driven Tests

The Problem with Hardcoded Data

CSV Data Set Config

Create Your Data File

Configure CSV Data Set

Use Variables in Requests

Sharing Modes Explained

All Threads (Default)

Current Thread Group

Current Thread

User Defined Variables

Environment-Specific Config

Complete E-commerce Data Setup

File Structure

users.csv

products.csv

Test Plan Configuration

Random Selection Functions

Random Number

Random from List

Random Product from CSV

Data Generation with Groovy

Parallel-Safe Data Strategies

Problem: Data Collisions

Solution 1: Unique Data Per Thread

Solution 2: Thread-Specific Data Files

Solution 3: Counter for Uniqueness

Key Takeaways

Next Lesson

🧠 Quick Quiz

What is the purpose of CSV Data Set Config?

What does 'Sharing mode: All threads' mean?

Why should you use 'Recycle on EOF: False' for user credentials?

Lesson Complete!

Get Your Free Resource

Start Your Free Trial

Your Interactive Session is Ready