Validate a Shopify Product CSV

Shopify's Body (HTML) column holds rich HTML. If a cell contains an unescaped comma and the export wasn't quoted, the row splits and every column shifts. The validator reports it as a row_width error with the exact column-count delta.

Header (6 cols):
Handle,Title,Body (HTML),Vendor,Variant Price,Variant SKU

Row (7 cols — one extra comma inside the HTML):
widget,Blue Widget,<p>Durable, compact frame</p>,Acme,12.99,WID-BLU
                                  ^ this comma was inside Body (HTML)
                                    but the cell was not quoted

Validator output:
Row 2, Col 1  (row width)
  type: row_width  severity: error
  detail: "Row has 7 columns; header has 6."

Fix: re-export from Shopify (Shopify quotes the field correctly), or in
     the spreadsheet wrap Body (HTML) cells in double quotes before saving.

Shopify expects bare decimals in Variant Price — no symbol, no thousands separator, no inline currency code. The validator infers the column as number and flags every cell that doesn't parse.

Header:
Handle,Variant Price,Variant Inventory Qty

Rows:
widget-blue,12.99,100        OK
widget-red,£14.99,50         type_mismatch  (currency symbol)
widget-green,15,40           OK  (whole numbers are fine)
widget-black,18,50 GBP,30    type_mismatch  (inline code + EU comma)

Validator output (2 issues):
Row 3: Expected number based on the column profile, but found "£14.99".
Row 5: Expected number based on the column profile, but found "18,50 GBP".

Fix: csv-find-replace regex [£$€¥] -> (empty), then fix EU comma decimals
     with (\d+),(\d+)$ -> $1.$2 before re-importing.

Two different products that share a Variant SKU cause Shopify to merge inventory tracking. Because the column name contains the sku token, it's checked as a unique key.

Header:
Handle,Title,Variant SKU,Variant Price

Rows:
blue-tee,Blue Tee,TEE-001,19.99
red-tee,Red Tee,TEE-001,19.99      <- same SKU, different product

Validator output:
Row 3, Col 3  Variant SKU
  type: duplicate_key  severity: warning
  detail: 'Value "TEE-001" also appears on row 2.
           This column looks like a unique key.'

Fix: give each variant a unique SKU, or de-dupe with csv-deduplicator on
     the Variant SKU column if the second row is a true duplicate.

Editing a Shopify export in Excel-for-Windows and saving as plain 'CSV (Comma delimited)' can write CP1252. The validator reads the file as UTF-8 (browser File.text() always decodes UTF-8), so CP1252 bytes for accented characters surface as the replacement character.

Header:
Handle,Title,Vendor

Rows (CP1252 bytes decoded as UTF-8):
cafe-mug,Caf\uFFFD Mug,Acme            encoding error (was 'Café')
creme-set,Cr\uFFFDme Brulee Set,Acme   encoding error (was 'Crème')

Validator output:
Row 2, Col 2 (Title): Cell contains a replacement or control character,
  which usually means an encoding problem.
Row 3, Col 2 (Title): same.

Fix: re-save from Excel as 'CSV UTF-8 (Comma delimited)', or re-export
     from Shopify (always UTF-8). The csv-cleaner tool can also normalise
     text before re-import.

When a column has only a handful of distinct values, the validator runs a suspicious-value scan: a value that appears once and is substring-similar to a value appearing two-plus times is flagged. Trailing-space typos in Status-style columns are the classic case.

Header:
Handle,Status

Rows (12 rows, Status column has 2 common values + 1 rare):
widget-1,active
widget-2,active
...   (8 more 'active' / 'draft' rows)
widget-12,active             <- trailing space, appears once
           ^ note the trailing space

Validator output:
Row 12, Col 2  Status
  type: suspicious_value  severity: warning
  detail: 'Rare value "active" looks close to common category "active".'

Fix: trim the trailing space with csv-whitespace-trimmer (or csv-find-replace),
     then re-validate so the value count collapses to the real two values.

Shopify's multi-row product format repeats the Handle across every variant of a product, so a 4-variant product has 4 rows sharing one Handle. Because Handle does not contain an id / sku / email token, the unique-key detector does not treat it as a key — so those legitimate repeats produce no duplicate_key warnings. Only the whole-row duplicate check could fire, and only if two variant rows are byte-for-byte identical. This is correct behaviour; do not try to de-dupe by Handle.

Tab/comma-delimited files can't hold a literal delimiter or hard newline inside a cell unless the cell is double-quoted. Rich-text Body (HTML) pasted from a CMS often carries both. The validator reports the affected row with row_width and the exact column-count delta. The cleanest fix is to re-export from Shopify (which quotes the field), or wrap Body (HTML) cells in double quotes; replacing the inner commas is a last resort because it edits your content.

On systems with a European locale, Excel often saves CSVs using semicolons as the field separator. PapaParse auto-detects the delimiter, so a semicolon file validates correctly as long as it's consistent. If a file mixes delimiters (some lines comma, some semicolon), the detector locks onto one and the other lines show up as row_width errors — that's the signal to re-export cleanly.

Many Shopify columns are legitimately blank on variant child rows or for products that don't use inventory tracking. The empty_cell warning only fires when a column is at least 85% populated, so a mostly-blank optional column stays quiet. If you do see empty-cell warnings on Variant Price, those are real — a priced column with a few blanks. Verify each warning corresponds to an intentionally-untracked variant before changing anything.

The number classifier strips a leading $ and removes commas before testing, so $1,299 parses as a number, but £1,299 (non-dollar symbol) does not and is flagged. EU comma-decimal like 12,99 also fails because the comma is stripped, leaving 1299, which is a different number than intended. The validator can only tell you the cell isn't a clean decimal — fix the formatting in source so the price is unambiguous.

The validator profiles only the columns that exist in the file. If you deleted the Variant Price column, the validator simply won't report on it — there is no schema to compare against. You are responsible for matching Shopify's expected header set. Open Shopify's sample product CSV, compare the header row, and add any missing columns before import.

Free tier is capped at 2 MB and 500 rows, where a row is a CSV line. Because Shopify spreads one product across multiple variant rows, even a modest catalog blows past 500 lines quickly (e.g. 170 products averaging 3 variants = ~680 lines). The runner blocks the file before validating and shows the limit message. Pro raises this to 100 MB / 100,000 rows; to validate a large file on free tier, split it with csv-row-splitter and validate each chunk.

The validator caps the reported issue list at 250 to keep the browser responsive, and the Issues tile reflects that capped count. If a file has a single systemic problem — wrong delimiter, a whole-column type drift, an encoding mismatch — you can easily exceed 250. Fix that one root cause and re-validate; the count usually drops to single digits. The downloaded JSON report contains the same capped issue list.

The validator reads text formats only (.csv, .tsv, .txt) via the browser's text decoder — it does not parse Excel binary workbooks. Dropping an .xlsx produces garbage rows or a parse failure, not a useful report. Save your workbook to CSV first (CSV UTF-8 recommended), then validate the saved file.