Import products into Adobe Commerce or Magento Open Source with a CSV

Which columns does the product attributes reference name?

The reference, last updated 8 October 2026, says each attribute is a column and each product is a row. Columns that begin with an underscore are service data. It says sku is the only required value, sku should be the first column, and the order of the other columns does not matter. A SKU can be up to 64 characters. A longer SKU fails the import. You can include only the columns you are changing. store_view_code identifies a store view, and a blank cell means the default store view. The page’s examples are storeview1, english, and spanish. attribute_set_code assigns the attribute set. The page’s example is default. After the product exists, the import cannot change the attribute set. product_type takes simple, grouped, configurable, virtual, or bundle, with the meanings printed on that page.

categories takes a path with a forward slash between a category and a subcategory. The same sentence says to separate multiple paths with a pipe, and it stops there. It does not print a pipe character. product_websites takes a website code, or several codes separated by a comma and no space. The page’s example is base or base,website2. name, description, and short_description are the product text. description can contain simple HTML. weight is the product weight. product_online uses 1 for enabled and 2 for disabled. tax_class_name is the tax class name. visibility uses the words Not Visible Individually, Catalog, Search, or Catalog, Search. price is the store price. special_price, special_price_from_date, and special_price_to_date cover the discounted range. url_key is the product part of the URL. save_rewrites_history uses 1 with a new url_key when you want a 301 from the old URL. The meta columns are meta_title, meta_keywords, and meta_description. The image columns named for a simple product include base_image, base_image_label, small_image, small_image_label, thumbnail_image, thumbnail_image_label, additional_images, and additional_image_labels. A new image path on that page is a file name with a leading slash, and the page’s example is /image.jpg. qty is the quantity in stock.

sku,product_type,attribute_set_code,name,price,product_online,qty
MUG-04,simple,default,Travel mug,18.00,1,12

Which statements about empty cells and categories disagree?

The CSV article, also last updated 8 October 2026, says a blank column is ignored during import, and that you cannot add an attribute that does not already exist. Its category example uses a comma between paths: Default Category/Gear,Default Category/Gear/Bags. It says the comma is the default multiple-value separator and that the import screen can use a different one. The attributes reference, on the categories row, says to separate multiple paths with a pipe. This page leaves those two sentences side by side. Match the separator to the Multiple value separator on the import screen you have open, and keep the header from the sample file.

The import article’s rule for existing products says empty values for non-required attributes replace the values already stored. The same article says Add/Update cannot remove a multiple-value field such as websites or categories by leaving it out of the file. Those values stay. A third sentence, on the CSV page, says blank columns are ignored. Do not treat an empty categories cell as a harmless blank until you know which of those three sentences your import is applying. For a large catalog, the import article says to omit categories and url_key when you are not changing them. If either column is present, Adobe Commerce rebuilds URL rewrites for every product in the file, and that work can time out on Adobe Commerce on Cloud.

How do you import the product CSV?

1. In the admin, go to System, then Data Transfer, then Import. Set Entity Type to Products and click Download Sample File. The sample has column headings and placeholder data for the product types. Keep that header. Advanced Pricing, customers, and Stock Sources are other entity types on the same screen, and each one is its own file.
2. In Golanto, choose Adobe Commerce / Magento. Record whether the shop is Adobe Commerce or Magento Open Source, the attribute set, and the date you downloaded the sample. The mapper still opens on Amazon until you change the platform. Map your values onto the sample’s headers. Put sku first. Write product_online as 1 or 2, which are the values the attributes reference prints.
3. Edit the CSV in a UTF-8 editor. The CSV article says Microsoft Excel inserts characters into the header that can stop the import, and that a Mac save should use CSV (Windows). The import article says a UTF-8 file with a byte-order mark can fail. A column that starts with an underscore is service data. A service name the shop does not have makes the file invalid.
4. Set Import Behavior to Add/Update, Replace, or Delete. Add/Update can change every product field except sku, creates new SKUs, and creates tax classes and categories that the file names. Replace deletes the existing product when the SKU matches and creates a new record. It errors when the SKU is not already in the database. Delete ignores every column except sku. The default field separator and multiple-value separator are commas. The empty-attribute constant defaults to _EMPTY_VALUE_. Check Data, then Import. The valid message is shown as a figure. This page does not quote it.
5. Images use the paths in the image columns. The import article says the default place to put image files on the server is pub/media/import. The same article says that from Adobe Commerce and Magento Open Source 2.3.2, Images File Directory is added to the images base directory var/import/images. Its example folder name is product_images. Use the path the message on your import screen describes. The default file-size limit printed in the article is 2 MB, and the screen’s own message shows the limit from php.ini.

Worked example

A supplier title column is not the Adobe Commerce name column

sku,title,price
MUG-04,Travel mug,18.00

These headings are a synthetic supplier sketch. The attributes reference names sku, name, and price. product_type simple and attribute_set_code default are values that page prints. product_online 1 means enabled, and 2 means disabled. sku stays first and within 64 characters. The sketch leaves out categories because the attributes page says to separate paths with a pipe and the CSV page’s example separates them with a comma. Copy the header from the sample file you downloaded, and set Multiple value separator to the character in that file.

Official references

• Adobe Commerce: Product data attributes reference (opens in a new tab)
• Adobe Commerce: CSV data files (opens in a new tab)
• Adobe Commerce: Import data (opens in a new tab)

Schema review: adobe-commerce-product-csv-2026-11-02. Examples are synthetic. Store-specific and unpublished requirements are outside these checks.

Common questions

Continue reading

• Map a product file to a marketplace template
• Review ambiguous prices and decimal separators
• See what OpenCart 4 documents about a product CSV
• Import products into Shopware with the Default product CSV