Sitemap Index

An INDEX, not a sitemap.

The catalogue is ~35,000 products served to 40 markets, and every market is a real URL with its own price and its own language. As one file that is 1.4 million entries — Google caps a sitemap at 50,000 URLs and 50 MB — so this lists one entry per market per chunk instead.

A sitemap index may NOT point at another index, so every chunk of every market is listed here directly: the hubs (sitemap-ie-1.xml, ...), then each offered tier of products in its own series (sitemap-ie-t1-1.xml, ...; see gcg/services/sitemap_tiers.py), so Search Console reports how much of each tier is indexed. Tier 1 alone is about 20,000 products, one part per market, so 40 markets are 80 entries, nowhere near the 50,000 an index may hold.

Each product part is dated by the newest lastmod it lists, read from the rows it is written from (_part_lastmod); a product's date is the same in every market, so one reading dates all forty copies. A hub part has no date to give: its entries carry none (see sitemap_chunk).

Counted from the same lists the parts are cut from (_SITEMAP_MEMO), so within an hour the index and the files it names agree exactly.