fix: COUNT(col) nulls in 3 partial-agg paths, COUNT(DISTINCT string) in SQL executor, PERCENTILE in isAggregate; cache LIKE regex

Author: teamchong

src/partial-agg.ts

@@ -186,8 +186,12 @@ export function computePartialAgg(
           updatePartialAgg(states[i], val, val);
         } else if (typeof val === "bigint") {
           updatePartialAgg(states[i], Number(val), val);
-        } else if (aggregates[i].fn === "count_distinct" && val !== null && val !== undefined) {
-          updatePartialAgg(states[i], 0, val);
+        } else if (val !== null && val !== undefined) {
+          if (aggregates[i].fn === "count_distinct") {
+            updatePartialAgg(states[i], 0, val);
+          } else if (aggregates[i].fn === "count") {
+            states[i].count++;
+          }
         }
       }
     }
@@ -219,8 +223,12 @@ export function computePartialAggColumnar(
             updatePartialAgg(states[i], val, val);
           } else if (typeof val === "bigint") {
             updatePartialAgg(states[i], Number(val), val);
-          } else if (aggregates[i].fn === "count_distinct" && val !== null && val !== undefined) {
-            updatePartialAgg(states[i], 0, val);
+          } else if (val !== null && val !== undefined) {
+            if (aggregates[i].fn === "count_distinct") {
+              updatePartialAgg(states[i], 0, val);
+            } else if (aggregates[i].fn === "count") {
+              states[i].count++;
+            }
           }
         }
       }
@@ -257,8 +265,12 @@ export function computePartialAggColumnar(
           updatePartialAgg(states[i], val, val);
         } else if (typeof val === "bigint") {
           updatePartialAgg(states[i], Number(val), val);
-        } else if (aggregates[i].fn === "count_distinct" && val !== null && val !== undefined) {
-          updatePartialAgg(states[i], 0, val);
+        } else if (val !== null && val !== undefined) {
+          if (aggregates[i].fn === "count_distinct") {
+            updatePartialAgg(states[i], 0, val);
+          } else if (aggregates[i].fn === "count") {
+            states[i].count++;
+          }
         }
       }
     }

src/sql/evaluator.ts

@@ -174,17 +174,24 @@ function toNumber(val: unknown): number {
 }
 
 /** Convert SQL LIKE pattern to case-insensitive match. % = any chars, _ = one char. */
+const _likeCache = new Map<string, RegExp>();
 function matchLike(value: string, pattern: string): boolean {
-  let re = "^";
-  for (let i = 0; i < pattern.length; i++) {
-    const ch = pattern[i];
-    if (ch === "%") re += ".*";
-    else if (ch === "_") re += ".";
-    else if (/[.*+?^${}()|[\]\\]/.test(ch)) re += "\\" + ch;
-    else re += ch;
+  let regex = _likeCache.get(pattern);
+  if (!regex) {
+    let re = "^";
+    for (let i = 0; i < pattern.length; i++) {
+      const ch = pattern[i];
+      if (ch === "%") re += ".*";
+      else if (ch === "_") re += ".";
+      else if (/[.*+?^${}()|[\]\\]/.test(ch)) re += "\\" + ch;
+      else re += ch;
+    }
+    re += "$";
+    regex = new RegExp(re, "i");
+    if (_likeCache.size > 64) _likeCache.clear();
+    _likeCache.set(pattern, regex);
   }
-  re += "$";
-  return new RegExp(re, "i").test(value);
+  return regex.test(value);
 }
 
 function castValue(val: unknown, targetType: string): unknown {
@@ -228,5 +235,6 @@ export function rewriteAggregatesAsColumns(expr: SqlExpr): SqlExpr {
 
 function isAggregate(name: string): boolean {
   return name === "COUNT" || name === "SUM" || name === "AVG" || name === "MIN" || name === "MAX" ||
-    name === "COUNT_DISTINCT" || name === "STDDEV" || name === "VARIANCE" || name === "MEDIAN";
+    name === "COUNT_DISTINCT" || name === "STDDEV" || name === "VARIANCE" || name === "MEDIAN" ||
+    name === "PERCENTILE";
 }

src/sql/executor.ts

@@ -196,9 +196,9 @@ function computeAgg(rows: Row[], agg: AggregateOp): number | bigint | string | b
     const v = row[agg.column];
     if (v === null || v === undefined) continue;
     const num = typeof v === "number" ? v : typeof v === "bigint" ? Number(v) : parseFloat(String(v));
+    if (agg.fn === "count_distinct") seen.add(String(v));
     if (!isNaN(num)) {
       values.push(num);
-      if (agg.fn === "count_distinct") seen.add(String(v));
     }
   }