I need to store only new records in a table

I need to create a table that stores only the sku_id and product_id columns that first appear in the published_products table with the date on which they first appeared. It should be noted that the sku_id are the different variants of a product_id. So far in bigquery it makes this query to obtain said calculation but the issue is that the products_pulicados table is partitioned by day and by not placing a partition date in the where condition within the query it makes me consume too many GB

    MIN(PARTITION_DATE) OVER(PARTITION BY sku_id) publication_date_variant,
    MIN(PARTITION_DATE) OVER(PARTITION BY product_id) publication_date_product,
  WHERE published = 1

I understand that I can create an incremental model but I can’t think of how I can make said model add only new records for the sku_id column with the date on which it appeared but also the information of the product_id column with the first time it appeared also said product.