Import a network export

The second file is your network's list of pending transactions. Export it as CSV, drop it on the Imports screen, and AffCheck turns it into rows it can match against your Shopify orders. This page covers what happens to the file; the page for your network covers where the export lives.

What happens when you drop a file

1. It is read here. The file is parsed in a Web Worker in this browser tab. There is no upload: the worker has no network access of its own.
2. The network is detected. AffCheck compares the header row with the layouts it knows and names the network it recognised. You can override it.
3. The rows are projected. Each row becomes ten fields — transaction id, order reference, date, publisher id, publisher name, voucher code, sale amount, commission, currency, status — plus the row number in your spreadsheet, so a flag can point you back at the line it came from.
4. They are sent in batches of two thousand and saved as transactions.
5. Matching is queued. As soon as the last batch lands, the matcher runs against your stored orders.

When it asks about columns

Every preset in AffCheck was written from each network's own documentation, and none of them has been confirmed against a real export yet. If your file's headers are not the ones the preset expects, AffCheck does not guess: it shows you the header row and asks which column is which.

It needs six: the order reference, the sale amount, the commission, the status, the transaction date and the publisher id. Four more are used when they exist: the transaction id, the publisher name, the voucher code and the currency.

The answer is saved against that network and that exact header layout, so you are asked once. Change your report's columns and you will be asked again — which is the point.

When it asks about currency

Some exports have no currency column. Rather than assume, AffCheck asks once which currency the file is in and uses your store's currency as the suggestion. Money is never converted between currencies anywhere in the product, so a file in the wrong currency produces a total you cannot reconcile — it is worth a second look at this question.

Importing the same file twice

Dropping the same file on the same network within a day is answered with "you already imported this" and the earlier import, because that is nearly always a mistake. "Import anyway" continues.

Importing it on purpose is safe. Rows are matched on the network's transaction id where the file has one, and on the order reference and transaction date where it does not, so a repeat updates the rows that are already there instead of adding a second copy. A failed import is never treated as a duplicate, so a retry always works.

What is stored

Only the projected rows. Not the file, not its name, not a column AffCheck did not read. The Imports list identifies an import by its network, the day it ran, how many rows it brought and who ran it.

After the import: matching

Matching does not run in the upload request; it runs in the background, and the card on the Imports screen tells you where it is. "Rows in, not matched yet" is a normal state for a few seconds. Once it finishes, the card shows rows, matched and flagged, and the Review screen has the flags.

If the count sits still, when an import fails or gets stuck explains each state. If the file was the wrong one, undo an import takes it back.

Your network