Skip to content

CSV feed file specification

Purpose

Uploaded files are stored as raw input and processed by the ETL pipeline to create or update product records.

File requirements

CSV files must meet the following requirements:

Requirement Description
File format CSV
Encoding UTF-8 recommended
Header row Required
Delimiter Comma ,
File extension .csv recommended
Content type text/csv, text/plain, or application/vnd.ms-excel
Required fields sku, product_name

Required columns

The following columns are required for each product record:

Column Type Required Description
sku string Yes Partner-provided unique product identifier
product_name string Yes Display name of the product

Rows missing sku or product_name are skipped during ETL processing.

The following columns are recommended for complete product records:

Column Type Required Description
brand string No Product brand or manufacturer
category string No Product category
price decimal No Product price (numeric, no currency symbol)
currency string No Three-letter currency code (for example, USD)
availability string No Product availability status
description string No Product description

Example CSV

sku,product_name,brand,category,price,currency,availability,description
SKU-1001,Wireless Mouse,Logitech,Electronics,24.99,USD,in_stock,Ergonomic wireless mouse
SKU-1002,USB-C Cable,Anker,Electronics,12.99,USD,in_stock,6-foot USB-C charging cable
SKU-1003,Laptop Stand,Generic,Office Accessories,39.99,USD,out_of_stock,Adjustable aluminum laptop stand

What happens

  • Validate CSV structure
  • Store the raw file in Amazon S3
  • Create submission and validation jobs
  • Process data during ETL
  • Insert, update, or skip product records

Validation

During ETL processing, the system validates each row.

Condition Result
Valid row with new sku Product is inserted
Valid row with existing sku and changed data Product is updated
Valid row with existing sku and no changes Product is unchanged
Row missing sku Row is skipped
Row missing product_name Row is skipped
Malformed CSV structure Validation may fail

Product uniqueness

Products are uniquely identified by the combination of:

partner_name + sku
This constraint prevents duplicate product ingestion for the same partner. This also means two different partners may use the same SKU without conflict.

Formatting guidelines

Follow these guidelines when preparing a CSV file:

  • Include exactly one header row
  • Do not leave required fields blank
  • Use consistent column names
  • Avoid duplicate sku values within the same partner feed
  • Use plain numeric values for prices, such as 19.99
  • Do not include currency symbols in the price field
  • Use consistent availability values, such as in_stock or out_of_stock