How to

Every table in your PDF, joined into one

Free converters give you one sheet per page. Mammoth reads the whole PDF: tables with the same columns come back as one table, even when one breaks across a page, and the headings and notes between them are left out.

46 seconds, real app, made-up sales report. Follow along with the same file: Regional sales report (PDF, 3 pages). Download the video (MP4)

The steps

1
Upload the PDF
In Data, choose New dataset and upload the PDF. When Mammoth asks how to read it, choose Extract Data.
2
Choose Only tables
The report has a table for each region, North, South and West, with headings and notes between them. South's table starts on page 1 and ends on page 2. Choose Only tables, then Extract.
3
Check what it found
Mammoth reports 1 table found, across pages 1 to 3, with 17 rows. The three tables share the same columns, so they came back as one, and the headings and the note were left out. Compare the row count with the PDF, then choose Looks right.
4
Create the dataset and explore it
Choose Create dataset. How the PDF was read is saved as the dataset's recipe, so next quarter's report is read the same way. In Explore, all 17 stores are in one table, with orders, revenue and returns as numbers you can total.

Keep the region on each row

Here the region is only in the headings above each table, so Only tables leaves it out. To keep it, choose Custom instead and say what you want, for example:

One row per store: region, store, manager, orders, revenue, returns

Mammoth fills the Region column from the heading each table sits under. Check it in the preview before you save.

If it doesn’t look right

The row count is lower than the PDF: Look at the preview for the rows next to a page break, and compare the first and last rows of each table with the PDF. If some are missing, read it again with Custom and describe the rows you want.
You got more than one table: Tables only come back as one when their columns match. If one region has an extra column, you'll see it as its own table. Use Custom to ask for one table with the columns you want.
A number came in as text: Check for footnote marks or currency symbols in the PDF. In Explore, change the column's type to Number; values that can't be read become blank, so check those rows.