marsdb 0.9.1

Embeddable property-graph database with an openCypher query subset: single binary, single file, optional in-memory mode.
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
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190

#![cfg(feature = "arrow")]
//! Arrow export — strict inference rules, exactness, batching. See
//! marsdb/src/arrow.rs's module docs for the rule table under test.

use std::collections::HashMap;

use arrow_array::{
    Array, BooleanArray, Date32Array, Float64Array, Int64Array, ListArray, RecordBatch, StringArray,
};
use arrow_schema::DataType;
use marsdb::{Database, ExecutionOptions};

fn arrow_all(db: &Database, cypher: &str, batch_rows: usize) -> Vec<RecordBatch> {
    db.query_arrow(
        cypher,
        &HashMap::new(),
        &ExecutionOptions::default(),
        batch_rows,
    )
    .unwrap()
    .collect::<Result<Vec<_>, _>>()
    .unwrap()
}

fn arrow_err(db: &Database, cypher: &str) -> String {
    db.query_arrow(cypher, &HashMap::new(), &ExecutionOptions::default(), 1024)
        .err()
        .expect("expected an Arrow export error")
        .to_string()
}

#[test]
fn scalar_types_map_and_round_trip_exactly() {
    let db = Database::in_memory().unwrap();
    db.execute(
        "CREATE (:N {i: 9223372036854775807, f: 1.5, s: 'héllo', b: true, \
         d: date('1984-10-11'), dur: duration('P1M2DT3H')})",
    )
    .unwrap();
    let batches = arrow_all(
        &db,
        "MATCH (n:N) RETURN n.i AS i, n.f AS f, n.s AS s, n.b AS b, n.d AS d, n.dur AS dur",
        1024,
    );
    assert_eq!(batches.len(), 1);
    let b = &batches[0];
    assert_eq!(b.num_rows(), 1);
    let schema = b.schema();
    assert_eq!(schema.field(0).data_type(), &DataType::Int64);
    assert_eq!(schema.field(1).data_type(), &DataType::Float64);
    assert_eq!(schema.field(2).data_type(), &DataType::Utf8);
    assert_eq!(schema.field(3).data_type(), &DataType::Boolean);
    assert_eq!(schema.field(4).data_type(), &DataType::Date32);
    assert!(matches!(schema.field(5).data_type(), DataType::Interval(_)));

    let ints = b.column(0).as_any().downcast_ref::<Int64Array>().unwrap();
    assert_eq!(ints.value(0), i64::MAX);
    let floats = b.column(1).as_any().downcast_ref::<Float64Array>().unwrap();
    assert_eq!(floats.value(0), 1.5);
    let strings = b.column(2).as_any().downcast_ref::<StringArray>().unwrap();
    assert_eq!(strings.value(0), "héllo");
    let bools = b.column(3).as_any().downcast_ref::<BooleanArray>().unwrap();
    assert!(bools.value(0));
    let dates = b.column(4).as_any().downcast_ref::<Date32Array>().unwrap();
    // 1984-10-11 = 5397 days since epoch.
    assert_eq!(dates.value(0), 5397);
}

#[test]
fn nulls_become_validity_not_type_changes() {
    let db = Database::in_memory().unwrap();
    db.execute("CREATE (:N {i: 1}), (:N), (:N {i: 3})").unwrap();
    let batches = arrow_all(&db, "MATCH (n:N) RETURN n.i AS i ORDER BY n.i", 1024);
    let col = batches[0].column(0);
    let ints = col.as_any().downcast_ref::<Int64Array>().unwrap();
    assert_eq!(ints.len(), 3);
    assert_eq!(ints.null_count(), 1);
}

#[test]
fn all_null_column_is_null_type_and_empty_result_works() {
    let db = Database::in_memory().unwrap();
    db.execute("CREATE (:N), (:N)").unwrap();
    let batches = arrow_all(&db, "MATCH (n:N) RETURN n.missing AS m", 1024);
    assert_eq!(batches[0].schema().field(0).data_type(), &DataType::Null);
    assert_eq!(batches[0].num_rows(), 2);

    let empty = arrow_all(&db, "MATCH (n:Nothing) RETURN n.x AS x", 1024);
    assert!(empty.is_empty());
}

#[test]
fn batching_splits_at_batch_rows() {
    let db = Database::in_memory().unwrap();
    for i in 0..10 {
        db.execute_with_params(
            "CREATE (:N {i: $i})",
            &HashMap::from([("i".to_string(), marsdb::PropertyValue::Int(i))]),
        )
        .unwrap();
    }
    let batches = arrow_all(&db, "MATCH (n:N) RETURN n.i AS i", 3);
    let sizes: Vec<usize> = batches.iter().map(|b| b.num_rows()).collect();
    assert_eq!(sizes, vec![3, 3, 3, 1]);
}

#[test]
fn homogeneous_lists_export_as_list_arrays() {
    let db = Database::in_memory().unwrap();
    db.execute("CREATE (:N {tags: [1, 2, 3]}), (:N {tags: [4]})")
        .unwrap();
    let batches = arrow_all(&db, "MATCH (n:N) RETURN n.tags AS tags", 1024);
    let col = batches[0].column(0);
    let lists = col.as_any().downcast_ref::<ListArray>().unwrap();
    assert_eq!(lists.len(), 2);
    let total: usize = (0..lists.len()).map(|i| lists.value(i).len()).sum();
    assert_eq!(total, 4);
}

#[test]
fn strict_errors_name_the_column() {
    let db = Database::in_memory().unwrap();
    db.execute("CREATE (a:N {x: 1})-[:R]->(b:N {x: 2.5})")
        .unwrap();
    // Mixed Int/Float in one column.
    let e = arrow_err(&db, "MATCH (n:N) RETURN n.x AS mixed");
    assert!(e.contains("'mixed'") && e.contains("Int and Float"), "{e}");
    // Whole entities.
    let e = arrow_err(&db, "MATCH (n:N) RETURN n");
    assert!(e.contains("project scalar properties"), "{e}");
    // Mixed list elements.
    let e = arrow_err(&db, "RETURN [1, 'a'] AS bad");
    assert!(e.contains("mixed list element types"), "{e}");
}

#[test]
fn stats_ride_along_and_options_apply() {
    let db = Database::in_memory().unwrap();
    let reader = db
        .query_arrow(
            "CREATE (:N {i: 1})",
            &HashMap::new(),
            &ExecutionOptions::default(),
            1024,
        )
        .unwrap();
    assert_eq!(reader.stats.nodes_created, 1);

    for _ in 0..5 {
        db.execute("CREATE (:N {i: 2})").unwrap();
    }
    let options = ExecutionOptions {
        max_result_rows: Some(2),
        ..Default::default()
    };
    assert!(db
        .query_arrow("MATCH (n:N) RETURN n.i", &HashMap::new(), &options, 1024)
        .is_err());
}

#[test]
fn ffi_stream_round_trips() {
    // Export through the C Data Interface stream and import it back --
    // the exact contract marsdb-capi and the PyCapsule protocol expose.
    use arrow_array::ffi_stream::{ArrowArrayStreamReader, FFI_ArrowArrayStream};

    let db = Database::in_memory().unwrap();
    db.execute("CREATE (:N {i: 9223372036854775807}), (:N {i: 2})")
        .unwrap();
    let reader = db
        .query_arrow(
            "MATCH (n:N) RETURN n.i AS i",
            &HashMap::new(),
            &ExecutionOptions::default(),
            1,
        )
        .unwrap();
    let mut ffi = FFI_ArrowArrayStream::new(Box::new(reader));
    let imported = unsafe { ArrowArrayStreamReader::from_raw(&mut ffi) }.unwrap();
    let batches: Vec<RecordBatch> = imported.collect::<Result<_, _>>().unwrap();
    assert_eq!(batches.len(), 2);
    let all: Vec<i64> = batches
        .iter()
        .flat_map(|b| {
            let a = b.column(0).as_any().downcast_ref::<Int64Array>().unwrap();
            (0..a.len()).map(|i| a.value(i)).collect::<Vec<_>>()
        })
        .collect();
    assert!(all.contains(&i64::MAX));
}