products.json must contain an array of rows (or { rows: [...] }) — mine is keyed by sku

Asked 17 Jun 20253,540 views2 answers

9votes

The export from our catalogue service is an object keyed by sku, which is the shape everything downstream of it expects:

json

{
  "SKU-1": { "name": "Backpack", "price": 29.99 },
  "SKU-2": { "name": "Bike light", "price": 9.99 }
}

Pointing a dataset at it gives:

text

✖ CONFIG_INVALID: /work/shop/projects/shop/data/common/products.json must contain an array of rows (or { rows: [...] }).

Odd part: sdods data preview on the same file does not complain at all. So one of the two is wrong about the file and I would like to know which.

asked by Koen Vermeulen309 · 17 Jun 2025

2 answers

8votes
Neither is wrong, they are lenient in different places, and preview is the lenient one.
A dataset is rows. The loader accepts two shapes: a top-level array, or an object with a rows array — the second so you can keep metadata beside the data without it becoming a column. A map keyed by sku is neither, hence the error.
Preview falls back to wrapping a bare object as a single row, so your keyed map previews as one row with a column per sku. If preview ever shows you 1 row(s) and a hundred columns, that is this shape problem and not a successful read.
Keep the key by promoting it to a column:
json
```
{
  "rows": [
    { "sku": "SKU-1", "name": "Backpack", "price": 29.99 },
    { "sku": "SKU-2", "name": "Bike light", "price": 9.99 }
  ]
}
```
Then Given I load dataset "products" where "sku" is "SKU-1" does what you meant. Same two shapes for a YAML source.
answered by Bruno Teixeira694 · 17 Jun 2025
3votes
If the export shape is not yours to change, we generate the dataset file from it in a small script and commit both. The transform is four lines and it means the failure happens at the transform, with a file name in it, rather than inside a scenario.
answered by Renata Kohl608 · 19 Jun 2025

Have a different question? Ask it.