sqrust_rules/lint/
insert_overwrite.rs

1use sqrust_core::{Diagnostic, FileContext, Rule};
2use std::collections::HashSet;
3
4pub struct InsertOverwrite;
5
6impl Rule for InsertOverwrite {
7    fn name(&self) -> &'static str {
8        "Lint/InsertOverwrite"
9    }
10
11    fn check(&self, ctx: &FileContext) -> Vec<Diagnostic> {
12        let source = &ctx.source;
13        let skip = build_skip_set(source);
14        let mut diags = Vec::new();
15
16        for (line, col) in find_two_word_keyword(source, "insert", "overwrite", &skip) {
17            diags.push(Diagnostic {
18                rule: self.name(),
19                message: "INSERT OVERWRITE is Hive/Spark SQL-specific syntax; use standard INSERT INTO or CREATE TABLE AS SELECT"
20                    .to_string(),
21                line,
22                col,
23            });
24        }
25
26        diags.sort_by_key(|d| (d.line, d.col));
27        diags
28    }
29}
30
31/// Builds a set of byte offsets that are inside string literals or line comments.
32fn build_skip_set(source: &str) -> HashSet<usize> {
33    let mut skip = HashSet::new();
34    let bytes = source.as_bytes();
35    let len = bytes.len();
36    let mut i = 0;
37    while i < len {
38        if bytes[i] == b'\'' {
39            i += 1;
40            while i < len {
41                if bytes[i] == b'\'' {
42                    if i + 1 < len && bytes[i + 1] == b'\'' {
43                        skip.insert(i);
44                        i += 2;
45                    } else {
46                        i += 1;
47                        break;
48                    }
49                } else {
50                    skip.insert(i);
51                    i += 1;
52                }
53            }
54        } else if i + 1 < len && bytes[i] == b'-' && bytes[i + 1] == b'-' {
55            while i < len && bytes[i] != b'\n' {
56                skip.insert(i);
57                i += 1;
58            }
59        } else {
60            i += 1;
61        }
62    }
63    skip
64}
65
66/// Finds occurrences of `word1` followed by whitespace and `word2` (case-insensitive),
67/// both at word boundaries, skipping positions in the skip set.
68/// Returns 1-indexed (line, col) of `word1` for each match.
69fn find_two_word_keyword(
70    source: &str,
71    word1: &str,
72    word2: &str,
73    skip: &HashSet<usize>,
74) -> Vec<(usize, usize)> {
75    let lower = source.to_lowercase();
76    let w1_len = word1.len();
77    let w2_len = word2.len();
78    let bytes = lower.as_bytes();
79    let len = bytes.len();
80    let mut results = Vec::new();
81    let mut i = 0;
82
83    while i + w1_len <= len {
84        if skip.contains(&i) {
85            i += 1;
86            continue;
87        }
88
89        if !lower[i..].starts_with(word1) {
90            i += 1;
91            continue;
92        }
93
94        // Word boundary before word1
95        let before_ok = i == 0
96            || {
97                let b = bytes[i - 1];
98                !b.is_ascii_alphanumeric() && b != b'_'
99            };
100
101        // Word boundary after word1
102        let after_w1 = i + w1_len;
103        let after_w1_ok = after_w1 < len && {
104            let b = bytes[after_w1];
105            !b.is_ascii_alphanumeric() && b != b'_'
106        };
107
108        if before_ok && after_w1_ok {
109            // Skip whitespace between word1 and word2
110            let mut j = after_w1;
111            while j < len && (bytes[j] == b' ' || bytes[j] == b'\t' || bytes[j] == b'\n' || bytes[j] == b'\r') {
112                j += 1;
113            }
114
115            if j + w2_len <= len && !skip.contains(&j) && lower[j..].starts_with(word2) {
116                // Word boundary after word2
117                let after_w2 = j + w2_len;
118                let after_w2_ok = after_w2 >= len
119                    || {
120                        let b = bytes[after_w2];
121                        !b.is_ascii_alphanumeric() && b != b'_'
122                    };
123
124                if after_w2_ok {
125                    let (line, col) = offset_to_line_col(source, i);
126                    results.push((line, col));
127                }
128            }
129        }
130
131        i += 1;
132    }
133
134    results
135}
136
137fn offset_to_line_col(source: &str, offset: usize) -> (usize, usize) {
138    let before = &source[..offset];
139    let line = before.chars().filter(|&c| c == '\n').count() + 1;
140    let col = before.rfind('\n').map(|p| offset - p - 1).unwrap_or(offset) + 1;
141    (line, col)
142}
sqrust_rules/lint/insert_overwrite.rs

sqrust_rules/lint/
insert_overwrite.rs
