How to Find Duplicate Image URLs in a Shopify Product CSV

A Shopify product export is a CSV where the Handle column is the product’s unique identifier — the slug from its storefront URL. One product is often several rows: the first row carries the product details and its first variant, and the Handle is then repeated on additional rows for extra variants and extra images.

The Image Src column holds each image’s URL, and Image Position is the order that image appears in the product gallery. An “extra image” row repeats the Handle, sets Image Src and Image Position, and leaves the other columns blank.

The first three rows are the three sizes of one hoodie, all pointing at one photo — the same Image Src repeated within a single Handle. The last two rows are two different products, cedar-desk-lamp and oak-desk-lamp, pointing at the same lamp photo — the same Image Src across two handles. ImageDup treats those two situations differently, and the section on DUPLICATE and REVIEW explains how.

Most repeats are a by-product of how the catalog was built and exported, not a deliberate choice:

• Variant rows. Sizes and colors of one product are separate rows. If the variants were never given their own photos, every row points at the same Image Src.
• Duplicated products. A new product is created by copying an existing one and the image URL is never changed, so a different product keeps the original’s photo.
• Placeholder images. A “coming soon” or generic category image is assigned to every product without a real photo yet. Dozens of unrelated handles can share one placeholder.
• Gallery rows. Extra-image rows for one product are normal and expected — they repeat the Handle, not usually the URL, but a bad export can duplicate a gallery entry.
• Migrations and re-exports. Moving between platforms, or exporting from a tool that appends rather than replaces, can put the same product and image on several rows.

A repeated URL is a signal to review, not automatically an error.

Checking a CSV before you import it? ImageDup works the same way: map the Handle column as the product identifier and Image Src as the image URL, then review repeated image addresses before uploading the file to Shopify.

The example Shopify-shaped CSV is built so each result type appears at least once:

• Intentional reuse. aurora-hoodie is listed on three rows, one per size, all pointing at one photo. Same handle, so ImageDup marks it REVIEW.
• A catalog mistake. cedar-desk-lamp and oak-desk-lamp are separate products pointing at the same lamp photo — a copy-paste error. Two handles, so it is a DUPLICATE.
• A shared placeholder. coming-soon.png is used by brass-hook-set, linen-napkin-set, and ceramic-planter. Three handles, so it is a DUPLICATE.
• Unique images. walnut-cutting-board has a two-row gallery with two different photos, and merino-scarf has a distinct photo per color. Neither is flagged.

ImageDup sorts every repeated Image Src into one of two groups, based on how many distinct Handle values share it:

• DUPLICATE — the URL is attached to two or more different, non-blank handles. Separate products are showing one picture, which is the case most likely to need a fix.
• REVIEW — the URL repeats, but the rows do not establish two or more different handles: one handle on several variant rows, a duplicated gallery row, or rows with a blank handle. Often a normal variant export; occasionally a data problem, so it is listed rather than ignored.

A row with a blank Handle never produces a DUPLICATE on its own, because ImageDup cannot tell whether it belongs to one product or several. This mirrors the tool’s behavior everywhere else; the matching rules on the homepage describe it in full.

Work through the flagged groups with three outcomes in mind:

• Keep. Variants of one product that genuinely look identical in the photo, or a bundle that reuses a component’s image on purpose. Nothing to change.
• Replace. A product pointing at another product’s image — the cedar-and-oak case. Give it its own Image Src.
• Review. Variant or gallery rows where you are not sure. Open the affected products in the Shopify admin before you change the export.

Do not bulk-delete rows from the CSV to “remove duplicates”. A row in a Shopify import is a variant or an image, not just a URL — deleting it can drop a size or a gallery photo. Correct the image reference and re-import, or make the change in the admin.

ImageDup normalizes an image URL conservatively before comparing: it trims surrounding whitespace, lowercases the scheme and host, and drops the #fragment. Everything else is kept exactly, because changing it could merge images that are actually different. So it will not group:

• Different ?v= values or size suffixes. Shopify’s CDN typically appends a ?v= version parameter and serves sizes like _grande or _1024x1024. The full query string is kept, so photo.jpg?v=1 and photo.jpg?v=2 stay separate, and photo_grande.jpg differs from photo_1024x1024.jpg.
• The same file on two hosts. A cdn.shopify.com URL and the same image on a custom domain or an origin server are different strings.
• Scheme, www, and trailing-slash differences. http versus https, www. versus the bare domain, and a trailing / are all treated as distinct.
• Path case. /Products/A.JPG and /products/a.jpg are not merged.

To catch the same picture served from different URLs, add a column of image hashes to your export and map it — see below.

ImageDup does not generate hashes, download images, or calculate anything from a picture. If your export already includes a column of hashes — an MD5 of each file, or a perceptual hash such as pHash or dHash, in hex — map it as Image hash. Matching then uses the hash for any row that has one and falls back to the Image Src URL for rows that do not. The hash values are trusted as given.

The main tool works with a product CSV from any platform, not just Shopify. The homepage also covers what counts as a duplicate and a longer guide to duplicate image URLs in product data.