XML product import into OpenCart without failures

Typical problems

• the import runs from the browser and stops at the PHP time limit;
• every update creates new products instead of updating existing ones;
• images are downloaded again every time and the disk fills with copies;
• products missing from the feed stay on sale with zero stock or, the other way round, get deleted along with their reviews and Google rankings;
• the site slows down for customers while the import runs.

How it should work

• run by cron from the command line, not from the browser, so there are no web server time limits;
• a stable product key: the supplier ID or SKU stored separately; products are matched only by it;
• only changed fields are updated: price and stock every time, description and images only when they change;
• products that leave the feed are disabled, not deleted, so the URL and reviews stay;
• category mapping between the supplier and the site is set once in a table;
• a lock against parallel runs so two imports don’t write to the database at once;
• a log: how many products were created, updated and disabled, and which errors happened.

Large feeds

A feed of tens of megabytes shouldn’t be loaded into memory whole. It’s read as a stream, product by product, and written to the database in batches, so the import takes minutes, not hours, and doesn’t get in customers’ way.

Several suppliers

With several suppliers each feed has its own format. The most reliable approach is to convert them all to one internal format and import with one piece of logic. A new supplier is then just a new “translator”, not a new module.

I’ve set up XML imports into OpenCart, images and galleries included, for several stores. If you need automatic catalog sync, it’s part of website integrations. If you need the opposite direction, exporting your products to marketplaces, see the free XML feed module.