# Changelog of Daijob Job Scraper – Bilingual Jobs in Japan & Asia (`blackfalcondata/daijob-com-scraper`) Actor

- **URL**: https://apify.com/blackfalcondata/daijob-com-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/blackfalcondata/daijob-com-scraper.md

## Changelog

All notable changes to this Actor are documented here.

### \[0.1.0] — 2026-11-18

Initial release.

#### Added

- daijob.com's own search filters as inputs: career level, posted by (employer, recruiter or staffing agency), company features (e.g. Remote Work), job category, industry, required Japanese level and minimum yearly salary in JPY - plus sorting by Japanese, English or Chinese level and by position level. A filter value the site does not offer stops the run with a clear input error instead of being ignored.
- New filters: posting language (English, Japanese or both boards), sort order, and "posted within N days". Leave the search term blank to browse the newest jobs.
- Locations like "Tokyo", "Japan" or "Singapore" now use daijob.com's own location filter instead of being added to the keywords; job page URLs can be pasted as start URLs, and pasted search URLs keep all their filters.
- Long incremental runs keep their claim on the saved search, and "Max results" now counts the jobs actually delivered.
- Runs stop on their own after 12 hours, or after 60 minutes in which no jobs are read; results collected before the stop are kept.
- A run where daijob.com cannot be read now fails clearly instead of reporting "no results"; Japanese-board cards read salary and language levels correctly.
- Scrape bilingual job listings from daijob.com by keyword or pasted search URL: title, company, salary, location, language levels and full job description.
- Every row now includes a direct link to the job page (`portalUrl`).
- Runs stop cleanly when the maximum cost per run set on the platform does not cover more results.
- Memory per run is capped at 2 GB.

#### Fixed

- With "posted within N days", a job whose posting date could not be read is no longer delivered (or charged) as a recent job; the run is marked partial instead.
- If daijob.com's location list cannot be read, a location like "Tokyo" is no longer silently turned into a keyword search; that search is skipped and the run is marked partial.
- A run that loses its claim on a saved search to another run stops delivering immediately and releases its lock, so the two runs never deliver the same jobs.
- A run that loses its saved search to another run no longer delivers (or charges) jobs it had already collected but not yet written; the other run delivers them.
- A briefly empty answer from daijob.com's location list is no longer remembered for the rest of the run; later searches for the same location try again.
- Compact output now keeps the record ID, job link and scrape time, so rows can be matched and de-duplicated across runs.
