Snowflake query here: SELECT u.identity as user_id, o.user_id as anonymous_id, o.time as timestamp, quantity_lookup.value as value FROM ( SELECT quantity_lookup.order_id, sum(items.value['quantity']) as value FROM ( SELECT * FROM ( SELECT * ,dense_rank() over (partition by order_number order by session_id desc) as rank FROM _2_CRO_CONVERSION_SHOPIFY_WEB_POST_CONFIRMED_ORDER WHERE time > '2021-11-01' ORDER BY time asc ) WHERE rank = 1 ) as quantity_lookup, table( flatten( input => try_parse_json(quantity_lookup.line_items) ) ) items GROUP BY quantity_lookup.order_id ORDER BY quantity_lookup.order_id desc ) as quantity_lookup LEFT OUTER JOIN _2_CRO_CONVERSION_SHOPIFY_WEB_POST_CONFIRMED_ORDER as o on o.order_id = quantity_lookup.order_id LEFT OUTER JOIN users as u on u.user_id = o.user_id

Awesome thank you

Will you let me know when that's done and I'll try to re-run?

If you're self-hosting, you'll need to pull the latest Docker container (app.growthbook.io updates automatically)

Just to close the loop for you, here you can see how this is being -- We're trying to understand how average basket sizes varies based on a price experiment we're running Basket size is not Revenue, but it is an order attribute -- You all don't have a metric concept to really capture this, but if I ignore the $ sign the revenue metric gets me there Food for thought on if you want to add a 'average of an event property' or something like that as a metric long term, but we're good for now!