joaoh82 / rust_sqlite / 25492531113
64%

DEFAULT BRANCH: main
Ran
Jobs 1
Files 51
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 64.728%. Remained the same
25492531113

push

github

web-flow 
feat(bench): Group A workloads — W2..W6 (SQLR-16, sub-phase 9.2) (#103)

Lands the OLTP-baseline workloads on top of the 9.1 harness:

  W2 range scan        WHERE secondary >= ? AND <= ?, three width
                       buckets (100/1k/10k). SQLRite full-scans
                       (no range index probe yet, see
                       supported-sql.md:210); SQLite uses the index.
  W3 bulk insert       BEGIN; INSERT..x100k; COMMIT in one txn.
                       iter_batched(PerIteration) so each criterion
                       sample starts from a fresh DB.
  W4 single-row        Auto-committed INSERTs against a 1k-row
                       preloaded table. Preload bounds SQLRite's
                       O(N) bottom-up rebuild at a stable table size.
  W5 mixed OLTP        50/50 SELECT-by-PK + UPDATE-by-PK on 100k
                       rows; even iters SELECT, odd iters UPDATE.
  W6 index lookup      WHERE secondary = ? on the unique secondary
                       index — both engines hit the index fast-path.

The plan calls for `BETWEEN x AND y` on W2; SQLRite doesn't ship
BETWEEN yet (`docs/supported-sql.md` "Not yet supported"), so the
workload uses the equivalent `>= AND <=` form. Bumping to BETWEEN
once the engine supports it = `W2.v2`.

Suite wiring:
  - benches/suite.rs grew per-workload register_w* fns sharing a
    db_path / tempdir_for helper. W2 fans into three criterion
    groups (one per width bucket) so each lands as its own
    BenchSample row in the JSON envelope.
  - W3's register_w3 is iter_batched + a per-sample counter so the
    SQLRite WAL sidecar from one sample doesn't collide with the
    next. sample_size(10) so a single-driver run finishes in
    minutes; override at the CLI for sharper estimates.
  - W4 setup pre-loads 1k rows in one BEGIN/COMMIT before the
    timed loop so per-iter cost reflects a stable table size,
    not a 0..N ramp.

Datasets:
  - data.rs gains GroupADataset (100k rows, secondary = uniq... (continued)

8704 of 13447 relevant lines covered (64.73%)

1.2 hits per line

Jobs
ID Job ID Ran Files Coverage
1 25492531113.1 51
64.73
 GitHub Action Run
Source Files on build 25492531113