Rehearse an import with representative catalog rows
Reconcile the ten synthetic records by identity, expected publication state and critical product fields before trusting an import summary. Keep comma handling, accented text and missing assets visible in the exception register. The table supplies test cases; obtain the current store template before turning them into an importable file.
Worked example — illustrative data.
| SKU | Synthetic source fields | Expected check |
|---|---|---|
| IMP-01 | Arc brass; USD 85; qty 3 | Identity, amount and stock |
| IMP-02 | Arc black; USD 85; qty 0 | Unavailable state |
| IMP-03 | Clamp lamp; USD 70; qty 2 | Collection membership |
| IMP-04 | Shade, cream; USD 20; qty 4 | Comma preserved in title |
| IMP-05 | Café cup; USD 18; qty 5 | Accented text preserved |
| IMP-06 | Box; USD 30; qty 1 | Dimensions mapped |
| IMP-07 | Small shirt; USD 25; qty 2 | Parent/variant relationship |
| IMP-08 | Medium shirt; USD 25; qty 2 | Distinct variant identity |
| IMP-09 | Manual item; blank image | Missing-data exception retained |
| IMP-10 | Retired sample; inactive | Publication state respected |
Treat these as test records, not a provider-valid CSV schema. Obtain the current import template from the actual store and map source fields to its supported columns. Preserve identifiers as text and check count, duplicates, amounts, currency, stock, images, categories and variant relationships after import. Reconcile the source sheet, import report and displayed product sample. Keep error rows in an exception register rather than silently dropping them.
Reconcile identities before and after reimport
Count distinct source product IDs and variant SKUs before importing the ten sample rows. After the first rehearsal, compare those identities, relationships and selected storefront states. Reimport one intentionally changed sample using the provider’s documented matching settings, then check whether it updated the intended item or created another. Preserve the import report and unmatched rows before attempting the full catalog.
Reimport matching behavior can create unintended duplicates
Reimporting a row may update or create an item depending on the configured matching behavior. Prove that behavior in staging before using the process on the live catalog.
Stop when counts, identity or variant relationships diverge. Correct the mapping and repeat the small rehearsal before scheduling a full import.
Does matching the total row count prove the import worked?
No. One missing variant and one duplicate can leave the total unchanged. Check stable identities, parent relationships, prices, images and publication states as well as counts. Keep an exception register for every unresolved mismatch.
Continue with Guides for stores, software teams and local businesses, eCommerce website design and development and Check ecommerce exports before migrating.
Before applying the specimen, verify exact feature availability, required permissions and supported field formats in the provider’s current documentation and your store.