There's not enough memory to complete this operation.

A query or refresh needed more RAM than was available — often a huge cross join, a fan-out relationship, or high-cardinality keys.

Why it happens

• Iterating very large tables, or CROSSJOIN/DISTINCT over high-cardinality columns.
• A relationship fan-out multiplying rows (a one-side key that isn't actually unique).
• Unnormalized keys — the same URL split into thousands of rows by query strings and casing.

How to fix it

1. Normalize keys before load: strip query strings, lowercase paths, map redirects to a stable id. This is often the single biggest reduction.
2. Avoid needless CROSSJOIN; filter earlier; prefer set-based aggregations to row-by-row iteration.
3. Check every relationship's one-side is genuinely unique.