workday has two separate bugs: totals cap at 2000 on affected queries, and any offset past the end wraps back to page 1. so add the loop guard to every workday pager
facet buckets within one facet should be mutually exclusive. overlap shows up across split queries or different facets so dedupe by a stable job key and trust only that count
also scope is companies on these ATSs with public boards. the bigger cost is maintenance, and remote/hybrid will probably take most of that effort because it isn't a standard field
Thanks, spot on. The actor handles both Workday quirks: when a board reports 2,000+ results it splits the search by a facet whose biggest bucket is under the cap, page offsets never go past min(total, 2000), and everything is deduplicated by the job's path. Remote/hybrid is exactly the maintenance cost you describe: each ATS's own field is mapped to remote/hybrid/onsite, with a fallback to the location text. Appreciate you reading it!
OnlineProxy
workday has two separate bugs: totals cap at 2000 on affected queries, and any offset past the end wraps back to page 1. so add the loop guard to every workday pager
facet buckets within one facet should be mutually exclusive. overlap shows up across split queries or different facets so dedupe by a stable job key and trust only that count
also scope is companies on these ATSs with public boards. the bigger cost is maintenance, and remote/hybrid will probably take most of that effort because it isn't a standard field