Structured vs Content Fields

When importing product data via a CSV import, I’ve previously blown up my database by tweaking the parent/child relationship column (I attempted to change it from using SKUs, by replacing them with IDs, which I figured was better practice). This resulted in many duplicated variable products – the newly created ones had no images, the old ones had no variations.

I have just again blown up my database in the CSV import. This time I was attempting to speed up the import process by marking various fields, such as Description and Attributes, as “Do not import”. They hadn’t changed, so I figured why import them again – just update the Price changes. When I subsequently checked my website, I no longer had variations, with just a message saying I first needed to add attributes to generate variations.

Trying to research this problem, I’ve come across the below, which would appear to explain the above issues:

Fields that are “structural” get wiped if present in the CSV but mapped to “Do not import.”
Fields that are “content” are ignored safely.

Structural fields include:

• Attributes
• Variation attributes
• Variation rows
• Parent/child relationships
• SKU (if blank)
• Product type (if blank)

Content fields include:

• Description
• Short description
• Regular price
• Sale price
• Stock quantity
• Weight
• Dimensions

Could someone who knows confirm if this is correct please. If so, it would suggest a user needs to be pretty careful when messing around with Structural Data.