Skip to content

How to

Next month's file, same steps

Rebuilding the same report every month? Clean the first file once. Next month, add the new file and every step replays. If the layout has changed, Mammoth stops and asks instead of breaking your data.

67 seconds, real app, made-up expenses. Follow along with the same three: October expenses (CSV), November expenses (CSV) and December, new layout (CSV). Download the video (MP4)

The steps

  1. 1

    Clean the first file once

    Upload October's file and open it in Explore. Add the steps it needs. Here, two: tidy the Category text into capital case, then rename Office Supplies to Office. Four clean categories are left.

    The pipeline with two steps, Text Formatting and Find & Replace on Category, and the Category column showing Meals, Software, Office and Travel
  2. 2

    Add next month's file

    When November's file arrives, select the dataset in Data and choose Add data. Upload the file and choose Submit. The new rows go on after the old ones.

  3. 3

    Check the new rows

    Open Explore. Both steps ran on the new rows without you touching them. November's file said TRAVEL, meals and office; it comes out as Travel, Meals and Office.

    November's rows in the table with clean categories: Meals, Travel, Office and Software
  4. 4

    If the layout changes, Mammoth stops

    December's file has an extra Project column, and Amount is now called Amount (GBP). Mammoth doesn't guess. It holds the file back and says so, and the dataset keeps the 26 rows it already had.

    A notice: 1 file has not been added. 1 file arrived and did not fit the table. The dataset still holds 26 rows from before. A Match the columns button
  5. 5

    Match the columns

    Choose Match the columns. Three of the four matched by name. Set Amount (GBP) against Amount. Project isn't in the table, so it's left out. Choose Apply changes.

    Column mapping with all four matched: Amount (GBP) to Amount, Date, Employee and Category, and a note that 1 incoming column is ignored
  6. 6

    December is in

    The file is added, 34 rows in all, and the same two steps cleaned it.

    The table with December's rows added, rows 27 to 34, with clean categories

Steps that replay well

Any step you add in Explore runs again, in order, on every file you add.

  • Tidying text: trim, capitals, fixing spellings
  • Renaming or merging values, such as Office Supplies to Office
  • Filtering out rows you never want, such as test entries
  • Joins to a lookup file, and calculated columns

Build the steps on a typical month, not an unusual one, so they fit the files that follow.

If it doesn’t look right

A new value slipped through
If next month has a spelling the steps don't cover, add it to the Find & Replace step. The fix applies to every month, old and new.
You uploaded the wrong file
Open Settings & history on the dataset. Each file you added is listed as its own batch there.
You'd rather replace the data than add to it
Add data can also swap the old file for the new one. Choose the pencil next to Update method in the Add data window before you submit.

Stop uploading by hand

On a paid plan, Automate new data adds files to the dataset as they land in a folder, so the steps run without you. On every plan, Add data works as shown here.
Add or replace data in the docs

Try it with two months of your own file.

Clean the first, add the second, and check the steps ran.

  • ✓ 14-day Pro trial
  • ✓ No credit card
  • ✓ Viewers always free