Data-Driven Tests

Data-driven tests let you run the same test logic with different input values. ProbeScript uses with examples: blocks with <variable> substitution.

Basic Example

test "login validation"
  open the app
  type <email> into "Email"
  type <password> into "Password"
  tap "Continue"
  see <expected>

with examples:
  email              password     expected
  "user@test.com"    "pass123"    "Dashboard"
  ""                 "pass123"    "Email is required"
  "user@test.com"    ""           "Password is required"

The test runs once for each row in the examples table. Each column header becomes a variable that can be referenced with <variable> syntax in the test body.

How It Works

1. The parser reads the with examples: block and identifies column headers from the first row
2. Each subsequent row creates a separate test instance
3. All <variable> references in the test body are replaced with the corresponding column value
4. Each instance is reported as a separate test result (e.g., “login validation [row 1]”, “login validation [row 2]“)

Column Formatting

Column values are separated by whitespace. Quoted strings can contain spaces:

with examples:
  name            email                 expected
  "Alice Smith"   "alice@test.com"      "Welcome, Alice"
  "Bob Jones"     "bob@test.com"        "Welcome, Bob"

Combining with Tags

Tags apply to all rows:

test "search results"
  @regression
  type <query> into "Search"
  tap "Go"
  see <result>

with examples:
  query          result
  "flutter"      "Flutter SDK"
  "dart"         "Dart Language"
  "probe"        "FlutterProbe"

Combining with Recipes

Data-driven tests work seamlessly with recipes:

use "recipes/auth.probe"

test "role-based access"
  log in as <email> with <password>
  see <page>

with examples:
  email               password     page
  "admin@test.com"    "admin123"   "Admin Panel"
  "user@test.com"     "user123"    "Dashboard"
  "guest@test.com"    "guest123"   "Read Only"

Loading Data from CSV

For large data sets, load examples from an external CSV file instead of inline tables:

test "login with CSV data"
  open the app
  type "<email>" into "Email"
  type "<password>" into "Password"
  tap "Sign In"
  see "<expected>"

with examples from "fixtures/users.csv"

The CSV file should have headers as the first row:

email,password,expected
user@test.com,pass123,Dashboard
admin@test.com,admin456,Admin Panel
guest@test.com,guest1,Limited Access

CSV paths are resolved relative to the .probe file’s directory. Absolute paths are also supported.

Random Data Generators

Use built-in generators to create random test data on each run:

test "registration with random data"
  open the app
  type "<random.email>" into "Email"
  type "<random.name>" into "Full Name"
  type "<random.phone>" into "Phone"
  type "<random.uuid>" into "Reference ID"
  type "<random.number(18,65)>" into "Age"
  type "<random.text(8)>" into "Invite Code"
  tap "Register"
  see "Success"

Available generators:

Generator	Example Output	Description
<random.email>	user_x7k2m@test.probe	Random email address
<random.name>	Alice Johnson	Random first + last name
<random.phone>	+1-555-042-7831	US-format phone number
<random.uuid>	550e8400-e29b-...	UUID v4
<random.number(min,max)>	42	Random integer in range
<random.text(length)>	aB3kM9xQ	Random alphanumeric string

Random generators are expanded before variable substitution, so they work in both inline and CSV-based data-driven tests.