1. Docs
  2. Resources
  3. Data Source
  4. Datasets (Excel / CSV)

Datasets: Excel and CSV tables

A dataset is a table of test data inside the project. Steps read its columns, a key can read one of its cells, and a pipeline can run a test once per row.

4 min read Updated CustomersAdmins

In short

On this page you can

  • Upload an XLSX, XLS or CSV file as a dataset, or start from an empty table.
  • Edit cells, add rows and delete rows right in the panel.
  • Read a column in any step with {{excel:name}}_column.
  • Let a pipeline run a test once per row of the dataset.
  • Keep a separate dataset per environment with a _TEST or _PRP name.

How to use it

  1. Open Datasets and click Upload dataset (Excel / CSV).
  2. Pick the file and tick Header if its first row holds column names.
  3. Check the name and the columns, then click Save.
  4. Use {{excel:name}}_column in a step, or point a key at one cell.
  5. In the pipeline, loop over the dataset to run once per row.

Upload a dataset#

  1. Open Datasets

    Go to Data SourceDatasets and click Upload dataset (Excel / CSV). To create it in a folder, use Add on that folder’s row.

  2. Pick the file

    Click Click to select a file (“Supported: XLSX, XLS, CSV”). For a workbook, only the first sheet is read. To type the data yourself, click or start with an empty table.

  3. Say whether the first row is a header

    Tick Header when the first row holds column names. Leave it off and the first row is data; the columns are then called column1, column2…

  4. Check the name and folder

    Name starts as the file name without its extension. This is the name steps use, so keep it short: users, products. Pick a Folder if you like.

  5. Save

    “Dataset created”. The dataset appears in the list with its size, its columns and Not used until a key reads it.

How column names are made#

Steps read columns by name, so the names must be safe to type. When you upload a file with Header ticked, each header is turned into a plain camelCase name without accents:

Header in the fileColumn name
Login EmailloginEmail
Müşteri AdımusteriAdi
passwordpassword
empty, a repeated name, or rowNocolumn + its position, for example column4

Blank lines in the file are skipped. Once saved, the column names stay exactly as stored, also when you edit and save again, so references in steps keep working.

Edit a dataset#

  • Open it: click its row in the list.
  • Change a cell: click it and type. “Click a cell to change it. # is the row number in a loop (_1, _2 …).”
  • Rows: Add row adds one at the end; Delete row on a row removes it. The table shows 50 rows per page (Previous / Next).
  • Replace all data: Replace from file. The dataset keeps its name. Before you save, Drop the file discards the picked file.
  • Save: “Unsaved changes — Save to keep them” shows until you click Save (“Dataset updated”).
  • Delete: Delete, then confirm “Are you sure you want to delete this record?” (“Dataset deleted”). Change keys that read a cell of it first; see The dataset could not be deleted.

Names are unique in the project: “This name already exists in this project. Please enter a different name.” Folders work as for connections: New folder, then Rename or Delete in the folder menu. Deleting a folder moves its datasets to the top level (“Are you sure you want to delete this folder? Configurations inside will move to the root.”).

Use a dataset in a test#

Every dataset is also a parameter named {{excel:name}}. The data source picker in Automate (tooltip “Data Sources - Global Parameters - Datasets”) inserts it for you.

You writeThe run gets
{{excel:users}}_emailColumn email of the current row in a loop; the first row when the test does not loop.
{{excel:users}}_#The number of rows.
A key bound to a cell, for example {{ds:standardUser}}Always that one cell. See Read a dataset cell.
Example · dataset users with header email, password
#  email                 password
1  ann@test.example.com  ••••
2  bob@test.example.com  ••••

Step "Type Email"    value  {{excel:users}}_email
Row 1 of the loop  → ann@test.example.com
Row 2 of the loop  → bob@test.example.com

Run once per row#

A test runs once per row when its pipeline loops over the dataset. You set this in the pipeline, in Schedule, with the execute count:

Execute countRows that run
{{excel:users}}_#All rows, in order
{{excel:users}}_#5The first 5 rows
{{excel:users}}_?All rows, in random order
{{excel:users}}_?55 random rows

In a pipeline that loops over a dataset, the test cases that read it run once per row; the others run once or N times. A Dry Run from Automate uses only the first row.

One dataset per environment#

Datasets follow the same rule as keys: a run in TEST that reads users uses users_TEST when that dataset exists, otherwise users. See Environment-specific keys.

Datasets are not secret

Every project member can open a dataset and see all its cells. Keep passwords out of datasets: put them in encrypted keys and keep only user names in the table.

If something goes wrong#

Didn’t find your answer?Write to support with your project name and what you see on screen.

info@smartestqa.ai
© 2026 Selenity Technology · SmartestQAsmartestqa.aiThis page as MarkdownChecked against smarttest-fe predev 2a8c79cc, smarttest-be 633c2a03, smarttest-testsuite (runner)