Why manual excel to json fails on real files

Operations sends an .xlsx, engineering needs JSON, and the conversion happens by exporting to CSV first — which is exactly where merged cells and leading zeros get destroyed.

You see this most often in handover files, imported catalogues, client data drops. Save-as-CSV flattens merged cells by repeating or dropping values, and converts ID columns to numbers.

What actually goes wrong

Manual handling fails in a specific way: it does not produce obviously broken output, it produces output that is almost right. Save-as-CSV flattens merged cells by repeating or dropping values, and converts ID columns to numbers. The mistake survives review because nothing looks wrong until an importer rejects the file or a report is quietly incorrect.

A cheaper habit

Check the file before it leaves your machine. That is the entire purpose of Excel to JSON — it reads the data you paste, applies the same rules an importer would, and shows the row that will break, rather than a summary of how many rows broke.

Because the workbook is parsed rather than converted to CSV first, an identifier column keeps the zeros that an export would strip.

For the full picture — imports, feeds, variants, identifiers — see the pillar guide: Shopify product data management.

Try it on your own file

Open Excel to JSON and paste a sample. It is free, it does not store anything, and you do not need an account. If the file is sensitive, save the page locally and run it offline.

Questions people ask

Is the file uploaded?

No — the workbook is parsed in the browser using the built-in decompressor.

What about merged cells?

The merged value appears once, in the top-left cell, and the rest of the row stays empty rather than shifted.