How to run a detail crawl
- Scrape the list page — search results, a catalogue, a directory.
- Choose Crawl these links… and pick the column that holds the links.
- Rowspell opens the first link. Pick the fields you want on that detail page.
- Start the crawl. Each link is visited in turn, and the detail fields are added as new columns.
Every value lands on its own row
Detail values are joined to rows by URL, never by position, so a value can only land on the row whose link produced it. The list keeps its order, and a Detail column says what happened on each row: read, nothing found, failed, refused, not read, or no link.
“Nothing found” means the page opened but did not match your fields; “failed” means the page did not load.
What counts toward a run
A crawl is a run of detail pages, and only a page that returns rows counts. Links from search engines are followed to where they really go before the page is read.
Save it
When the crawl starts from a recipe, the detail fields are saved with it, so the two-level scrape can be run again.