PMM-13807 fingerprint keywords

go.mod

@@ -1,6 +1,6 @@
 module github.com/percona/go-mysql
 
-go 1.24.2
+go 1.23.4
 
 require (
 	github.com/pkg/errors v0.9.1

go.sum

@@ -1,6 +1,5 @@
 github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
-github.com/go-ole/go-ole v1.2.6 h1:/Fpf6oFPoeFik9ty7siob0G6Ke8QvQEuVcuChpwXzpY=
 github.com/go-ole/go-ole v1.2.6/go.mod h1:pprOEPIfldk/42T2oK7lQ4v4JSDwmV0As9GaiUsvbm0=
 github.com/go-ole/go-ole v1.3.0 h1:Dt6ye7+vXGIKZ7Xtk4s6/xVdGDQynvom7xCFEdWr6uE=
 github.com/go-ole/go-ole v1.3.0/go.mod h1:5LS6F96DhAwUc7C+1HLexzMXY1xGRSryjyPPKW6zv78=

query/query.go

@@ -728,13 +728,13 @@ func Fingerprint(q string) string {
 		default:
 			if s != inWord && s != inOp {
 				// If in a word or operator then keep copying the query, else
-				// previous chars were being ignored for some reasons but now
+				// previous chars were being ignored for some reason, but now
 				// we should start copying again, so set cpFromOffset.  Example:
 				// col=NOW(). 'col' will be set to copy, but then '=' will put
 				// us in inOp state which, if a value follows, will trigger a
 				// copy of "col=", but "NOW()" is not a value so "N" is caught
 				// here and since s=inOp still we do not copy yet (this block is
-				// is not entered).
+				//  not entered).
 				if Debug {
 					fmt.Println("Random character")
 				}
@@ -766,12 +766,21 @@ func Fingerprint(q string) string {
 			copy(f[fi:fi+l], prevWord)
 			fi += l
 			cpFromOffset = cpToOffset
-			if wordIn(prevWord, "in", "value", "values") && sqlState != onDupeKeyUpdate {
-				// IN ()     -> in(?+)
-				// VALUES () -> values(?+)
+			valueKeywords := []string{"in", "value", "values"}
+			if wordIn(prevWord, valueKeywords...) && sqlState != onDupeKeyUpdate {
 				addSpace = false
-				s = inValues
-				sqlState = inValues
+				// since we only consider the keywords if it exists at a top-level (except onDupeKeyUpdate),
+				// we only do the conversion below if the "keywords" actually open a value list (i.e., before opening a parenthesis).
+				if r == '(' || (isSpace(r) && (q[qi+1] == '(') || (q[qi+1] == ' ')) {
+					// IN ()     -> in(?+)
+					// VALUES () -> values(?+)
+					s = inValues
+					sqlState = inValues
+				} else {
+					f[fi] = ' '
+					fi++
+					cpFromOffset++
+				}
 			} else if addSpace {
 				if Debug {
 					fmt.Println("Add space")

query/query_test.go

@@ -165,6 +165,41 @@ func TestFingerprintBasic(t *testing.T) {
 			query:    "select * from foo limit 5 offset 10",
 			expected: "select * from foo limit ? offset ?",
 		},
+		{
+			name:     "insert with quoted keyword as column",
+			query:    "INSERT INTO test (ID, `Value`) VALUES (1 ,1)",
+			expected: "insert into test (id, `value`) values(?+)",
+		},
+		{
+			name:     "insert with non-quoted keyword as column",
+			query:    "INSERT INTO test (ID, Value) VALUES (1 ,1)",
+			expected: "insert into test (id, value) values(?+)",
+		},
+		{
+			name:     "insert with non-quoted keyword and space as column",
+			query:    "INSERT INTO test (ID, Value ) VALUES (1 ,1)",
+			expected: "insert into test (id, value ) values(?+)",
+		},
+		{
+			name:     "insert with keyword as first column name",
+			query:    "INSERT INTO test ( In , ID) VALUES (1,1)",
+			expected: "insert into test ( in , id) values(?+)",
+		},
+		{
+			name:     "insert with value keyword as first column name",
+			query:    "INSERT INTO test ( In , ID) VALUES (1,1)",
+			expected: "insert into test ( in , id) values(?+)",
+		},
+		{
+			name:     "insert duplicate with keyword as column name",
+			query:    "INSERT INTO test (id, value) VALUES (1, 10) ON DUPLICATE KEY UPDATE value = VALUES(value) + 5",
+			expected: "insert into test (id, value) values(?+) on duplicate key update value = values(value) + ?",
+		},
+		{
+			name:     "insert duplicate with keyword and space as column name",
+			query:    "INSERT INTO test (id, value ) VALUES (1, 10) ON DUPLICATE KEY UPDATE value = VALUES(value ) + 5",
+			expected: "insert into test (id, value ) values(?+) on duplicate key update value = values(value ) + ?",
+		},
 		{
 			name:     "fingerprint load data infile",
 			query:    "LOAD DATA INFILE '/tmp/foo.txt' INTO db.tbl",