Batch File – Adjustments and Cancellations

This guide walks you through how to report order adjustments and cancellations to Button via a batch file upload.

While Button's preferred integration path for reporting new orders, adjustments, and cancellations is via our Orders API,, we also support reporting these via a batch file. This guide walks through how to setup reporting via batch file – specifically, adjustments and cancellations.

General Guidelines

  • Button will provide you with a dedicated SFTP directory for uploading your batch files
  • For the first ever batch file upload, the file should contain all the order adjustment and cancellation events from the previous interval (e.g. for a 24 hour cycle it should be yesterdays events)
  • Subsequently, batch files should capture all adjustment and cancellation events that occurred since the previous upload (e.g. for a 24 hour cycle the previous day from 00:00:00 UTC to 23:59:59 UTC)
  • You cannot create an order via this process
  • The entire new order state after the adjustment/cancellation should be reflected in the batch file
  • If there is an order in the batch file that cannot be matched to an original order in Button’s systems (based on the order ID), it will be ignored
  • If the order in the batch file matches to one already in the Button system (based on the order ID), but has been finalized or cancelled already, it will be ignored

Batch File Format

Batch files should be delivered in a comma-separated value (CSV) format and the file name should specify the date & time it was generated. For example, if on a 24 hour cadence, BRANDNAME_orders_2019-01-01.csv.

Each line item in the order should have a row in the batch file (i.e. if there are 3 items in the order, there should be 3 rows: 1 for each line item).

There should be only one row for each updated_date, order_id, line_item_identifier combination. If this combination appears multiple times in the batch file, we will accept the row with the latest updated_date.

Immutable Data

The following fields cannot be changed from the original order. If they are included in the batch file, that entire column will be ignored.

  • currency
  • customer_order_id
  • purchase_date
  • btn_ref
  • customer_id
  • customer_email_sha256
  • customer_advertising_id

Reporting Adjustments

When adjusting an order, you must provide the current state of the order rather than the delta. If line items that are present in the original order are not included in the adjustment file, they will be removed from the order.

All of these scenarios must be accounted for:

  • Line Item quantity increase
  • Line Item quantity decrease
  • Line Item total increase
  • Line Item total decrease
  • Line Item addition
  • Line Item removal

Sample CSV

action,updated_date,order_id,line_item_identifier,line_item_total,line_item_quantity,line_item_sku,line_item_upc,line_item_category,line_item_subcategory,line_item_description
ADJUST,2019-04-30T00:04:32Z,OI-1024,LI-2048,199,1,1234-XCYZ,100000299906,Grocery,Fruits,Fuji Apples
ADJUST,2019-04-30T00:04:32Z,OI-1024,LI-512,420,6,2405-CDGF,100000249906,Grocery,Fruits,Bananas
ADJUST,2019-04-30T15:45:00Z,OI-4096,LI-256,998,1,1704-SNGV,100000249904,Grocery,Drinks,Oat Milk

Required Data for Adjustments

All of these columns must be present in the batch file with valid data for the order adjustment to be processed:

Column NameDescriptionData Type
actionThe type of update being applied to the line itemString (accepted inputs: ADJUST, NONE)
updated_dateDate representing the time the order was updatedISO-8601 string
order_idThe order id unique to this order that was passed to Button when the order was created.ASCII string (max 255 characters)
line_item_identifierThe unique identifier for this line item, within the scope of this order. This must be unique across all line-items within the order. We suggest using the SKU or UPC of the productString
line_item_totalTotal price of all items bought in a particular line item (e.g. if 3 medium pizzas were purchased for $10.00 each, total would be 3000); represented in the smallest currency unit (e.g., 100 for $1.00) in the same currency as the initial orderInteger > 0
line_item_quantityThe number of unique units represented by this line item. Defaults to 1Integer > 0

Optional Data

Column NameDescriptionData Type
line_item_skuThe Stock Keeping Unit of the line item. A Merchant-specific identifier for the line itemString (max 255 characters)
line_item_upcThe Universal Product Code of the line item. Also referred to as UPC-A and GTIN-12. UPC-E (8-digit) codes should be converted to 12-digit codesString (12 numeric digits)
line_item_categoryThe top-most category of the line itemString
line_item_subcategoryThe second-level category of the line item (i.e. immediately below the topmost category)String
line_item_descriptionText describing the line itemString

The above batch file when processed will attempt to adjust two orders in Button's systems:

  1. Order with ID OI-1024 which will now have two line items with IDs LI-2048 and LI-512
  2. Order with ID OI-4096 which will now have one line item with ID LI-256

Reporting Cancellations

If any of the line items in an order has a “CANCEL” action, the entire order will be cancelled.

Cancellations are final and cannot be reversed.

Required Data for Cancellations / Invalidations

All of these columns must be present in the batch file with valid data for the order cancellation to be processed:

Column NameDescriptionData Type
actionThe type of update being applied to the line itemString; accepted inputs: CANCEL, NONE
updated_dateDate representing the time the order was updatedISO-8601 string
order_idThe order id unique to this order that was passed to Button when the order was created.ASCII string (max 255 characters)

Sample CSV

action,updated_date,order_id
CANCEL,2019-04-30T00:04:32Z,OI-1024
CANCEL,2019-04-30T15:45:00Z,OI-4096

The above batch file when processed will attempt to cancel two orders in Button's systems:

  1. Order with id OI-1024 will be cancelled completely
  2. Order with id OI-4096 will be cancelled completely