All files server/explain/duckdb-parser.ts

93.22% Statements 55/59
93.87% Branches 46/49
100% Functions 3/3
92.85% Lines 52/56

Press n or j to go to the next uncovered block, b, p or k for the previous block.

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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192                                                                      208x 208x 208x       208x   208x   432x 404x     392x 392x   380x 188x   380x 44x       380x 4x       380x     380x 76x     380x   208x     172x 172x 112x   172x     380x       208x                 208x                       380x 380x 1404x 1024x   380x     380x               392x         392x                     384x 384x   4x       380x       380x                       380x 380x 392x 392x 392x       392x     44x 44x 44x 44x     380x               380x    
/**
 * DuckDB EXPLAIN output parser
 * DuckDB uses a similar format to PostgreSQL but with some differences
 * Parses text format EXPLAIN output and normalizes to common structure
 */
 
import type { ExplainOperation, ExplainResult, ExplainSummary } from '../types/executor'
 
/**
 * Parse DuckDB EXPLAIN output
 *
 * Example DuckDB EXPLAIN output:
 * "┌───────────────────────────┐"
 * "│      EXPLANATION OF       │"
 * "│     QUERY PLAN            │"
 * "└───────────────────────────┘"
 * "┌─────────────────────────────────────────────────────────────────────┐"
 * "│                     QUERY PLAN                                      │"
 * "├─────────────────────────────────────────────────────────────────────┤"
 * "│  HASH_JOIN                                                          │"
 * "│  ├──SEQ_SCAN employees                                              │"
 * "│  │  (cost=100.0 rows=1000)                                          │"
 * "│  └──SEQ_SCAN departments                                            │"
 * "│     (cost=50.0 rows=500)                                            │"
 * "└─────────────────────────────────────────────────────────────────────┘"
 *
 * Or simpler format:
 * "HASH_JOIN"
 * "├──SEQ_SCAN employees"
 * "└──SEQ_SCAN departments"
 */
export function parseDuckDBExplain(
  rawOutput: string[],
  sqlQuery: { sql: string; params?: unknown[] }
): ExplainResult {
  const operations: ExplainOperation[] = []
  const usedIndexes: string[] = []
  let hasSequentialScans = false
  let totalCost: number | undefined
 
  // Stack for building hierarchical structure
  const stack: { indent: number; op: ExplainOperation }[] = []
 
  for (const line of rawOutput) {
    // Skip decorative lines (box drawing characters)
    if (/^[┌├└│─┐┤┘]+$/.test(line.trim())) continue
    if (/EXPLANATION|QUERY PLAN/i.test(line)) continue
 
    // Parse operation lines
    const operation = parseDuckDBOperationLine(line)
    if (operation) {
      // Track sequential scans and indexes
      if (operation.type.includes('SEQ_SCAN') || operation.type.includes('TABLE_SCAN')) {
        hasSequentialScans = true
      }
      if (operation.type.includes('INDEX_SCAN') && operation.index) {
        usedIndexes.push(operation.index)
      }
 
      // Track total cost (from root operation)
      if (operations.length === 0 && operation.estimatedCost !== undefined) {
        totalCost = operation.estimatedCost
      }
 
      // Calculate indentation level (count tree characters)
      const indent = countTreeIndent(line)
 
      // Pop stack until we find a parent with less indentation
      while (stack.length > 0 && stack[stack.length - 1].indent >= indent) {
        stack.pop()
      }
 
      if (stack.length === 0) {
        // Root level operation
        operations.push(operation)
      } else {
        // Child operation
        const parent = stack[stack.length - 1].op
        if (!parent.children) {
          parent.children = []
        }
        parent.children.push(operation)
      }
 
      stack.push({ indent, op: operation })
    }
  }
 
  const summary: ExplainSummary = {
    database: 'duckdb',
    planningTime: undefined,
    executionTime: undefined,
    totalCost,
    hasSequentialScans,
    usedIndexes: [...new Set(usedIndexes)],
  }
 
  return {
    operations,
    summary,
    raw: rawOutput.join('\n'),
    sql: sqlQuery,
  }
}
 
/**
 * Count indentation level based on tree drawing characters
 */
function countTreeIndent(line: string): number {
  let indent = 0
  for (const char of line) {
    if (char === ' ' || char === '│' || char === '├' || char === '└' || char === '─') {
      indent++
    } else {
      break
    }
  }
  return indent
}
 
/**
 * Parse a single DuckDB EXPLAIN line
 */
function parseDuckDBOperationLine(line: string): ExplainOperation | null {
  // Remove tree drawing characters and trim
  const trimmed = line
    .replace(/[┌├└│─┐┤┘]/g, '')
    .replace(/^\s*/, '')
    .trim()
 
  if (!trimmed) return null
 
  // Match DuckDB operation pattern
  // Examples:
  // "HASH_JOIN"
  // "SEQ_SCAN employees"
  // "INDEX_SCAN idx_employees_org on employees"
  // "FILTER (organisation_id = 'org-1')"
  // "(cost=100.0 rows=1000)"
 
  // Check for cost line
  const costMatch = trimmed.match(/^\(cost=([\d.]+)\s+rows=(\d+)\)$/i)
  if (costMatch) {
    // This is a cost annotation, not an operation
    return null
  }
 
  // Match operation with optional table/index
  const opMatch = trimmed.match(
    /^([A-Z_]+)(?:\s+(\S+))?(?:\s+on\s+(\S+))?(?:\s+\(cost=([\d.]+)\s+rows=(\d+)\))?/i
  )
 
  Iif (!opMatch) {
    // Check for filter pattern
    const filterMatch = trimmed.match(/^FILTER\s+(.+)$/i)
    if (filterMatch) {
      return {
        type: 'FILTER',
        filter: filterMatch[1],
      }
    }
    return null
  }
 
  const type = opMatch[1].toUpperCase()
  let table = opMatch[2] || undefined
  let index = opMatch[3] || undefined
  const cost = opMatch[4] ? parseFloat(opMatch[4]) : undefined
  const rows = opMatch[5] ? parseInt(opMatch[5], 10) : undefined
 
  // For INDEX_SCAN format: "INDEX_SCAN <index_name>" or "INDEX_SCAN <index_name> on <table_name>"
  // opMatch[2] is the index name, opMatch[3] (after "on") is the table name
  if (type === 'INDEX_SCAN') {
    // Swap: what was captured as 'table' is actually the index name
    // and what was captured as 'index' (after "on") is actually the table name
    const actualIndex = table
    const actualTable = index
    index = actualIndex
    table = actualTable
  }
 
  const operation: ExplainOperation = {
    type,
    table,
    index,
    estimatedRows: rows,
    estimatedCost: cost,
  }
 
  return operation
}