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
use crate::manifest::{DataFile, FileContent, Snapshot};
use crate::storage::Storage;
use anyhow::Result;
/// A single unit of work for reading a table.
/// Combines a data file with its relevant delete files.
#[derive(Debug, Clone)]
pub struct ScanTask {
pub data_file: DataFile,
pub delete_files: Vec<DataFile>,
}
/// A simple predicate for data pruning.
#[derive(Debug, Clone)]
pub enum Predicate {
/// Equality predicate: column_id == value
Eq { column_id: i32, value: Vec<u8> },
/// Set membership: column_id IN (values)
In {
column_id: i32,
values: Vec<Vec<u8>>,
},
}
/// Plan scans for a table snapshot.
pub struct ScanPlanner<'a> {
snapshot: &'a Snapshot,
storage: &'a Storage,
filter: Option<Predicate>,
}
impl<'a> ScanPlanner<'a> {
pub fn new(snapshot: &'a Snapshot, storage: &'a Storage) -> Self {
Self {
snapshot,
storage,
filter: None,
}
}
/// Adds a filter to the scan planner.
pub fn with_filter(mut self, filter: Predicate) -> Self {
self.filter = Some(filter);
self
}
/// Plans the scan by associating data files with relevant delete files.
pub async fn plan(&self) -> Result<Vec<ScanTask>> {
let (data_files, delete_files) = self.snapshot.all_files(self.storage).await?;
// Group delete files by type
let mut pos_deletes = Vec::new();
let mut eq_deletes = Vec::new();
for df in delete_files {
match df.content {
FileContent::PositionDeletes => pos_deletes.push(df),
FileContent::EqualityDeletes => eq_deletes.push(df),
_ => {}
}
}
// For this prototype, we'll associate all equality deletes with all data files
// and filter position deletes by file path if we were to read them here.
// In a real implementation, we'd use partition pruning for deletes too.
let tasks = data_files
.into_iter()
.filter(|df| self.should_keep_file(df))
.map(|data_file| {
// Find relevant position deletes for this specific file.
// (Simplified: In a real system we'd use metadata to avoid searching all)
let mut relevant_deletes = Vec::new();
// Add all equality deletes (conservative)
relevant_deletes.extend(eq_deletes.clone());
// Add all position deletes (reader will filter)
relevant_deletes.extend(pos_deletes.clone());
ScanTask {
data_file,
delete_files: relevant_deletes,
}
})
.collect();
Ok(tasks)
}
fn should_keep_file(&self, data_file: &DataFile) -> bool {
if let Some(ref filter) = self.filter {
match filter {
Predicate::Eq { column_id, value } => {
if let Some(stats) = data_file.statistics.get(column_id) {
if let Some(ref bf) = stats.bloom_filter {
return bf.contains(value);
}
}
}
Predicate::In { column_id, values } => {
if let Some(stats) = data_file.statistics.get(column_id) {
if let Some(ref bf) = stats.bloom_filter {
return values.iter().any(|v| bf.contains(v));
}
}
}
}
}
true
}
}