Changes for version 0.009.0 - 2026-10-12
- Enhancements
- aggregate(): group the merged view by one or more published column names and compute aggregate measures (sum/avg/min/max/count/count_distinct) over each group. Supports HAVING, sort_by, limit and offset. On the sqlite/auto backend the GROUP BY is pushed into the existing SQLite spill cache so rows never pass through Perl; the array backend folds merged rows in Perl. The auto backend always takes the SQLite path for aggregation (skips the row-count threshold probe, which only exists to decide whether to spill — an aggregation query never fetches every row). Without this method a joined view could not answer "total per group" without shipping every merged row to a caller.
- _sqlite_criteria, _sqlite_where, _sqlite_join_clause: extract the criteria partition, WHERE builder and JOIN builder from _sqlite_join as Protected helpers so _sqlite_aggregate can reuse them without duplicating logic.
Documentation
Modules
Read-only combined view across two or more Database::Abstraction objects