From 73078bf73870be23142d3c6703d52b5da3279d17 Mon Sep 17 00:00:00 2001
From: Alan Protasio <alanprot@gmail.com>
Date: Tue, 30 May 2023 04:49:22 -0700
Subject: [PATCH] Opmizing Group Regex (#12375)

Signed-off-by: Alan Protasio <alanprot@gmail.com>
---
 tsdb/querier.go            |  9 ++++++++-
 tsdb/querier_bench_test.go |  4 ++++
 tsdb/querier_test.go       | 21 +++++++++++++++++++++
 3 files changed, 33 insertions(+), 1 deletion(-)

diff --git a/tsdb/querier.go b/tsdb/querier.go
index 9baf3f242..72b6b5141 100644
--- a/tsdb/querier.go
+++ b/tsdb/querier.go
@@ -190,7 +190,14 @@ func findSetMatches(pattern string) []string {
 	}
 	escaped := false
 	sets := []*strings.Builder{{}}
-	for i := 4; i < len(pattern)-2; i++ {
+	init := 4
+	end := len(pattern) - 2
+	// If the regex is wrapped in a group we can remove the first and last parentheses
+	if pattern[init] == '(' && pattern[end-1] == ')' {
+		init++
+		end--
+	}
+	for i := init; i < end; i++ {
 		if escaped {
 			switch {
 			case isRegexMetaCharacter(pattern[i]):
diff --git a/tsdb/querier_bench_test.go b/tsdb/querier_bench_test.go
index 19619e35b..c6deaeb44 100644
--- a/tsdb/querier_bench_test.go
+++ b/tsdb/querier_bench_test.go
@@ -113,6 +113,7 @@ func benchmarkPostingsForMatchers(b *testing.B, ir IndexReader) {
 	jXplus := labels.MustNewMatcher(labels.MatchRegexp, "j", "X.+")
 	iCharSet := labels.MustNewMatcher(labels.MatchRegexp, "i", "1[0-9]")
 	iAlternate := labels.MustNewMatcher(labels.MatchRegexp, "i", "(1|2|3|4|5|6|20|55)")
+	iNotAlternate := labels.MustNewMatcher(labels.MatchNotRegexp, "i", "(1|2|3|4|5|6|20|55)")
 	iXYZ := labels.MustNewMatcher(labels.MatchRegexp, "i", "X|Y|Z")
 	iNotXYZ := labels.MustNewMatcher(labels.MatchNotRegexp, "i", "X|Y|Z")
 	cases := []struct {
@@ -132,6 +133,7 @@ func benchmarkPostingsForMatchers(b *testing.B, ir IndexReader) {
 		{`j=~"XXX|YYY"`, []*labels.Matcher{jXXXYYY}},
 		{`j=~"X.+"`, []*labels.Matcher{jXplus}},
 		{`i=~"(1|2|3|4|5|6|20|55)"`, []*labels.Matcher{iAlternate}},
+		{`i!~"(1|2|3|4|5|6|20|55)"`, []*labels.Matcher{iNotAlternate}},
 		{`i=~"X|Y|Z"`, []*labels.Matcher{iXYZ}},
 		{`i!~"X|Y|Z"`, []*labels.Matcher{iNotXYZ}},
 		{`i=~".*"`, []*labels.Matcher{iStar}},
@@ -161,6 +163,8 @@ func benchmarkPostingsForMatchers(b *testing.B, ir IndexReader) {
 
 	for _, c := range cases {
 		b.Run(c.name, func(b *testing.B) {
+			b.ReportAllocs()
+			b.ResetTimer()
 			for i := 0; i < b.N; i++ {
 				_, err := PostingsForMatchers(ir, c.matchers...)
 				require.NoError(b, err)
diff --git a/tsdb/querier_test.go b/tsdb/querier_test.go
index 5bf721a62..e9dd3b75f 100644
--- a/tsdb/querier_test.go
+++ b/tsdb/querier_test.go
@@ -2051,6 +2051,12 @@ func TestPostingsForMatchers(t *testing.T) {
 				labels.FromStrings("n", "2"),
 			},
 		},
+		{
+			matchers: []*labels.Matcher{labels.MustNewMatcher(labels.MatchNotRegexp, "n", "(1|2.5)")},
+			exp: []labels.Labels{
+				labels.FromStrings("n", "2"),
+			},
+		},
 		{
 			matchers: []*labels.Matcher{labels.MustNewMatcher(labels.MatchEqual, "n", "1"), labels.MustNewMatcher(labels.MatchNotRegexp, "i", "^a$")},
 			exp: []labels.Labels{
@@ -2112,6 +2118,13 @@ func TestPostingsForMatchers(t *testing.T) {
 				labels.FromStrings("n", "1", "i", "b"),
 			},
 		},
+		{
+			matchers: []*labels.Matcher{labels.MustNewMatcher(labels.MatchRegexp, "i", "(a|b)")},
+			exp: []labels.Labels{
+				labels.FromStrings("n", "1", "i", "a"),
+				labels.FromStrings("n", "1", "i", "b"),
+			},
+		},
 		{
 			matchers: []*labels.Matcher{labels.MustNewMatcher(labels.MatchRegexp, "n", "x1|2")},
 			exp: []labels.Labels{
@@ -2134,6 +2147,14 @@ func TestPostingsForMatchers(t *testing.T) {
 				labels.FromStrings("n", "2.5"),
 			},
 		},
+		{
+			matchers: []*labels.Matcher{labels.MustNewMatcher(labels.MatchRegexp, "i", "(c||d)")},
+			exp: []labels.Labels{
+				labels.FromStrings("n", "1"),
+				labels.FromStrings("n", "2"),
+				labels.FromStrings("n", "2.5"),
+			},
+		},
 	}
 
 	ir, err := h.Index()