# Northmeld: clean a file and reuse the recipe

This short test uses fictional customer exports. All names, companies, and contact details are sample data. Valid sample emails use example.com. You do not need your own customer data.

## What you need

- Northmeld: https://www.northmeld.com/
- An account on the live site. Local CSV/Excel cleanup is free.
- Both CSV files from this kit, saved to your computer:
  - `01-fictional-customers-september.csv`
  - `02-fictional-customers-october.csv`

Local files stay in your browser by default. Optional cloud features are separate and are not needed for this test. Use the same browser and account throughout so you can access the recipe you save.

Import the original CSV files directly. Opening and saving them in another spreadsheet application first may change the intended test data.

## 1. Import the September file

Sign in and open the workspace. Choose **Open file** and select `01-fictional-customers-september.csv`.

Confirm that the table has 21 data rows and these eight columns:

`CustomerID`, `Name`, `Email`, `Company`, `Country`, `SignupDate`, `MonthlySpendUSD`, `Status`.

## 2. Preview and apply the cleanup

Open **Review cleaning plan** and enable these five operations. Leave other operations off for this first test:

- Trim whitespace.
- Remove empty rows.
- Remove duplicate rows.
- Lowercase email addresses.
- Normalize safe dates, using `YYYY-MM-DD` as the output format.

Before applying, look at a few before/after values and the rows proposed for removal. Is it clear what will happen?

Apply the plan. The table should now have **18 data rows**: two repeated full rows and one empty row have been removed.

The two people named **Alex River** should both remain. They have different customer IDs and emails. Missing emails, invalid emails, uncertain dates, and questionable amounts still need human review; this cleanup does not supply missing facts.

## 3. Try undo and redo

Undo the cleanup once. Confirm the original values and 21 data rows return. Redo it once to restore the cleaned table with 18 rows.

Keep the cleaned result active before saving the recipe.

## 4. Save the recipe

Open **Recipes**, choose **Save current**, and name it `Monthly customer cleanup`. Choose **Save reusable recipe**.

Save the recipe before making any row-specific manual corrections. This test is about reusing the five cleanup operations.

## 5. Apply the recipe to the October file

Use **Open file** to import `02-fictional-customers-october.csv` as a new table rather than appending it to September.

Open **Recipes**, find `Monthly customer cleanup`, and choose **Run on current file**. If field mapping is requested, match each column to the same name and inspect any compatibility warning before continuing.

The October table should also finish with **18 data rows**. Both records named **Jordan Glen** should remain. You should not need to configure the five cleanup operations again.

## 6. Review the remaining issues and export

Check these intentionally unresolved values in each file:

| Issue | September customer | October customer |
| --- | --- | --- |
| Missing email | C09011 | C10011 |
| Email missing the @ symbol | C09012 | C10012 |
| Email with two @ symbols | C09013 | C10013 |
| Missing country | C09014 | C10014 |
| Uncertain day/month order | C09015: 03/04/2026 | C10015: 04/05/2026 |
| Amount recorded as text | C09017: not recorded | C10017: not recorded |
| Unusually high amount requiring review | C09018: 9800 | C10018: 12500 |

The zero amounts at C09016 and C10016 are intentional and should remain zero. The company names containing a comma should remain in one cell, and the Chinese names should remain readable.

Country variants such as US, USA, and United States are also intentional. The five operations above do not standardize these labels. You can optionally try a separate explicit field mapping after completing the main test.

Try exporting the October result as CSV. If export requires review, inspect the remaining issues and follow the review controls. Do not invent missing information to make the warnings disappear. If you cannot proceed, report the step and message you encountered.

## 7. Reply with three things

Please reply in the Reddit thread where you found this kit:

1. Could you tell which values would change and which rows would be removed before applying the plan?
2. Did saving the recipe and applying it to the second file work? Where did you get stuck, if anywhere?
3. What kind of CSV or Excel exports do you normally clean, and would this workflow help?

A short reply is enough. Screenshots are optional; please do not include private data, account details, or passwords. If something failed, include the step number and the error message.