fsqlite 0.1.18

Public API facade
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74

//! Profile-first: which aggregate/existence shapes still SCAN (Rewind) vs SEEK after the recent
//! eq/range/IN/composite-prefix-range work? Run:
//! cargo test -p fsqlite --test gap_probe2 -- --ignored --nocapture

use fsqlite::Connection;
use fsqlite_types::SqliteValue;

fn shape(conn: &Connection, sql: &str) -> (bool, bool) {
    let ops: Vec<String> = conn
        .query(&format!("EXPLAIN {sql}"))
        .unwrap()
        .iter()
        .filter_map(|r| match r.values().get(1) {
            Some(SqliteValue::Text(op)) => Some(op.to_string()),
            _ => None,
        })
        .collect();
    (
        ops.iter().any(|o| o.starts_with("Seek")),
        ops.iter().any(|o| o == "Rewind"),
    )
}

#[test]
#[ignore = "probe; run with --ignored --nocapture"]
fn gap2() {
    let c = Connection::open(":memory:").unwrap();
    for ddl in [
        "CREATE TABLE t (id INTEGER PRIMARY KEY, a INTEGER, b INTEGER, x INTEGER);",
        "CREATE INDEX idx_ab ON t(a, b);",
        "CREATE INDEX idx_a ON t(a);",
    ] {
        c.execute(ddl).unwrap();
    }
    for i in 1..=300_i64 {
        c.execute(&format!(
            "INSERT INTO t VALUES ({i}, {}, {}, {});",
            i % 15,
            i % 30,
            i % 50
        ))
        .unwrap();
    }
    let cases = [
        "SELECT MIN(b) FROM t WHERE a = 7 AND b > 2",
        "SELECT MAX(b) FROM t WHERE a = 7 AND b > 2",
        "SELECT COUNT(*) FROM t WHERE a = 7 AND b > 3",
        "SELECT COUNT(*) FROM t WHERE a = 7 AND x = 5",
        "SELECT EXISTS(SELECT 1 FROM t WHERE a = 7 AND b = 3)",
        "SELECT a, COUNT(*) FROM t GROUP BY a",
        "SELECT MIN(b) FROM t WHERE a = 7",
        "SELECT AVG(b) FROM t WHERE a = 7 AND b > 3",
        "SELECT COUNT(*) FROM t WHERE a = 7 AND b BETWEEN 3 AND 9",
        "SELECT MIN(a) FROM t WHERE a > 5",
        // Pure single-column range aggregates (no residual): guarded by agg_range_shadowed_index_oracle.
        "SELECT COUNT(*) FROM t WHERE a < 5",
        "SELECT COUNT(*) FROM t WHERE a >= 5 AND a <= 9",
        // Still-scanning shapes worth a future lever: GROUP BY, COUNT(DISTINCT), bare MIN/MAX walk.
        "SELECT COUNT(DISTINCT a) FROM t",
        // Non-aggregate single-column range scans (codegen_select paths 1906/2088) — do these seek
        // the index under the composite-shadow schema, or shadow-fail like the aggregate range did?
        "SELECT id FROM t WHERE a < 5",
        "SELECT id FROM t WHERE a < 5 ORDER BY a",
        "SELECT id, b FROM t WHERE a >= 5 AND a <= 9",
        "SELECT id FROM t WHERE a >= 5 AND a <= 9 ORDER BY a",
    ];
    eprintln!("\n########## gap2: SEEK vs SCAN ##########");
    for sql in cases {
        let (seek, rewind) = shape(&c, sql);
        let tag = if seek { "SEEK " } else { "SCAN!" };
        eprintln!("  [{tag}] seek={seek} rewind={rewind}  {sql}");
    }
    eprintln!("########## end ##########\n");
}