CSV CHECK
Find encoding, price, stock, and duplicate issues before importing a STORES product CSV.
🔒 Analysis runs in your browser. The CSV file itself is not uploaded to BlueprintLab.
| Severity | Row | Column | Issue | Value |
|---|
Checks the STORES item bulk-import CSV for deterministic format and field issues.
Check encoding, price, stock and duplicate values before import.
アイテム名,価格,紹介文,追加在庫,バリエーション,品番,バーコード,原価,カテゴリ,画像1 Sample T-shirt,2500,Sample item,10,M,TS-M,4901234567890,1200,Tops,sample.jpg
Use this before a STORES bulk item-registration CSV upload to check encoding, CSV structure, header shape, required values, price, stock, variations, and identifiers together. Since the feature can register up to 300 items at once, it is much faster to isolate suspicious rows before scanning hundreds by eye.
STORES bulk item registration is for adding new items, not bulk-updating existing items. Confirm the workflow itself before debugging field values.
Current STORES guidance for item bulk registration requires Shift_JIS. If the file is UTF-8, fix the save format before spending time on individual cells. The same applies to a modified or unrecognized header: return to the current sample format first.
The checker also surfaces the CSV size and 300-item limits early. If either limit is the blocker, splitting the file matters more than fixing row-level warnings.
Encoding, CSV syntax, and header shape are file-level problems. Item name, price, stock, SKU, barcode, and similar findings are row-level problems. Fix file-level errors first; otherwise you can spend time cleaning rows in a file that still cannot be uploaded.
Some STORES business rules—such as whether a variation should exist for a particular item—cannot be decided from primitive types alone. Finish with the official sample and a small real registration.
For 300 items, build a representative ten-item file using the same columns. Verify Shift_JIS, header structure, and row values, run a real registration, then scale the working structure to the remaining items.
A ten-row end-to-end test is usually faster than discovering a template problem after uploading all 300 rows.