Speed up result counting iterator

[mattmundell]

What:

Speed up iterator created by init_result_get_iterator_severity.

Tests indicate about a 4x increase in result counting speed (for report counts cache).

Why:

Faster.

How did you test it:

On master:

1. time gvmd --optimize rebuild-report-cache => 280m
2. alter table report_counts rename to report_counts_master;
3. drop index report_counts_by_report_and_override;
4. CREATE TABLE IF NOT EXISTS report_counts (id SERIAL PRIMARY KEY, report integer REFERENCES reports (id) ON DELETE RESTRICT, "user" integer REFERENCES users (id) ON DELETE RESTRICT, severity decimal, count integer, override integer, end_time integer, min_qod integer);
5. Switch to patch.
6. time gvmd --optimize rebuild-report-cache => 77m
7. select count(*) from report_counts; => 21280
8. select count(*) from report_counts_master; => 21280
9. select report, user, severity, count, override, end_time, min_qod from report_counts_master EXCEPT select report, user, severity, count, override, end_time, min_qod from report_counts; => 0 rows

Checklist:

• Tests
• CHANGELOG Entry