Cross-project alignment of dependency updates: A reproducible mining pipeline across five ecosystems and four PR sources

Dataset deposit for the paper Cross-project alignment of dependency updates: A reproducible mining pipeline across five ecosystems and four PR sources. This record contains the pseudonymised release SQLite database and JSON-Schema column documentation used to reproduce the paper's released analyses. The SQLite database contains per-triple alignment statistics, change-level classification labels, and human-verified gold-set PR/change labels. It intentionally excludes raw PR metadata tables, changed-file lists, raw parsed change rows, PR titles, PR bodies, closing-comment bodies, commit-message text, plaintext GitHub logins, commit author emails, and the hashing salt. Public dependency package identifiers and version pairs are retained only at the aggregate triple level. Go module import paths are released as public package/module names after checking that all observed Go modules were publicly listed by the Go module proxy or had public version metadata. The accompanying schema.json file documents every released table and column.