Use datasets for planned input and runtime values for captured output. That one rule keeps suites readable and dramatically reduces confusion during debugging.
EDITING A DATASET
- Datasets are plain key/value files (CSV or JSON) stored inside your project - readable in any editor, versionable in git.
- In the app, the Dataset Editor shows Key / Value columns with Add, Delete, Save, and Save As... A value can be marked Encrypted for secrets you do not want stored in plain text.
- In a suite step, set Value Mode to dataset and put the KEY in the value field. At run time the key resolves to its dataset value.
THE TWO NAMESPACES
- Dataset keys: planned before the run. usernames, passwords, labels, postal codes, expected values.
- Runtime variables: captured during the run by capture steps with a Save As name, reused later as ${RUNTIME_NAME}.
- Never use the same name for both. Prefix if it helps: EXPECTED_CART_COUNT (dataset) vs RUNTIME_CART_COUNT (captured).
THE DEBUGGING TIP THAT SAVES THE MOST TIME
If a field gets typed empty during a run, the dataset key is almost always missing or misspelled. Key names must match exactly - including hyphens vs underscores (track-input and track_input are different keys). Open the dataset, confirm the exact key exists, and re-run. The same rule applies to expected values in assertions.
Need the detailed workflow?
The public guide stays intentionally high level. Open the Intelligent Assistant for click-by-click instructions, learning order, and practice ideas.
Open Assistant