feat(SQLNormalizer): apply suggestions

Eldolfin · Eldolfin · commit 61d01a2c30bd · 2026-04-27T12:00:17.000+02:00
diff --git a/internal-api/src/main/java/datadog/trace/api/normalize/SQLNormalizer.java b/internal-api/src/main/java/datadog/trace/api/normalize/SQLNormalizer.java
@@ -114,6 +114,9 @@ private static BitSet findSplitterPositions(byte[] utf8) {
       if (b == '\'' && !escaped && !doubleQuoted) {
         singleQuoted = !singleQuoted;
       } else if (b == '"' && !singleQuoted) {
+        // No !escaped check here: in SQL, double quotes delimit identifiers,
+        // where backslash has no special escaping role. A literal double quote
+        // inside an identifier is escaped by doubling it (""), not by backslash.
         doubleQuoted = !doubleQuoted;
       } else {
         escaped = (b == '\\') & !escaped;
diff --git a/internal-api/src/test/groovy/datadog/trace/api/normalize/SQLNormalizerTest.groovy b/internal-api/src/test/groovy/datadog/trace/api/normalize/SQLNormalizerTest.groovy
@@ -130,11 +130,11 @@ INSERT INTO country (country_name, country_name_eng, country_code) VALUES (?, ?,
 INSERT INTO country (country_name, country_name_eng, country_code) VALUES (?, ?, ?);"""
     "SELECT * FROM TABLE WHERE userId = ',' and foo=foo.bar" | "SELECT * FROM TABLE WHERE userId = ? and foo=foo.bar"
     "SELECT * FROM TABLE WHERE userId =     ','||foo.bar" | "SELECT * FROM TABLE WHERE userId =     ?||foo.bar"
-    "CLOSE \"<unnamed portal 1>\""                        | "CLOSE \"<unnamed portal 1>\""
-    "CLOSE \"<unnamed portal 1>\\\""                      | "CLOSE \"<unnamed portal 1>\\\""
-    "FETCH ALL IN \"<unnamed portal 1>\""                 | "FETCH ALL IN \"<unnamed portal 1>\""
-    "SELECT 1 WHERE x = \"col 1\""                        | "SELECT ? WHERE x = \"col 1\""
-    "SELECT \"x\\\", 123 FROM t"                          | "SELECT \"x\\\", ? FROM t"
+    /CLOSE "<unnamed portal 1>"/         | /CLOSE "<unnamed portal 1>"/
+    /CLOSE "<unnamed portal 1>\"/        | /CLOSE "<unnamed portal 1>\"/
+    /FETCH ALL IN "<unnamed portal 1>"/  | /FETCH ALL IN "<unnamed portal 1>"/
+    /SELECT 1 WHERE x = "col 1"/         | /SELECT ? WHERE x = "col 1"/
+    /SELECT "x\", 123 FROM t/            | /SELECT "x\", ? FROM t/
     "SELECT count(*) AS totcount FROM (SELECT \"c1\", \"c2\",\"c3\",\"c4\",\"c5\",\"c6\",\"c7\",\"c8\", \"c9\", \"c10\",\"c11\",\"c12\",\"c13\",\"c14\", \"c15\",\"c16\",\"c17\",\"c18\", \"c19\",\"c20\",\"c21\",\"c22\",\"c23\", \"c24\",\"c25\",\"c26\", \"c27\" FROM (SELECT bar.y AS \"c2\", foo.x AS \"c3\", foo.z AS \"c4\", DECODE(foo.a, NULL,NULL, foo.a ||', '|| foo.b) AS \"c5\" , foo.c AS \"c6\", bar.d AS \"c1\", bar.e AS \"c7\", bar.f AS \"c8\", bar.g AS \"c9\", TO_DATE(TO_CHAR(TO_DATE(bar.h,'YYYYMMDD'),'DD-MON-YYYY'),'DD-MON-YYYY') AS \"c10\", TO_DATE(TO_CHAR(TO_DATE(bar.i,'YYYYMMDD'),'DD-MON-YYYY'),'DD-MON-YYYY') AS \"c11\", CASE WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,'YYYYMMDD'))) > 150 THEN '>150 Days' WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,'YYYYMMDD'))) > 120 THEN '121 to 150 Days' WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,'YYYYMMDD'))) > 90 THEN '91 to 120 Days' WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,'YYYYMMDD'))) > 60 THEN '61 to 90 Days' WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,'YYYYMMDD'))) > 30 THEN '31 to 60 Days' WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,'YYYYMMDD'))) > 0 THEN '1 to 30 Days' ELSE NULL END AS \"c12\", DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,'YYYYMMDD')),NULL) as \"c13\", bar.k AS \"c14\", bar.l ||', '||bar.m AS \"c15\", DECODE(bar.n, NULL, NULL,bar.n ||', '||bar.o) AS \"c16\", bar.p AS \"c17\", bar.q AS \"c18\", bar.r AS \"c19\", bar.s AS \"c20\", qux.a AS \"c21\", TO_CHAR(TO_DATE(qux.b,'YYYYMMDD'),'DD-MON-YYYY') AS \"c22\", DECODE(qux.l,NULL,NULL, qux.l ||', '||qux.m) AS \"c23\", bar.a AS \"c24\", TO_CHAR(TO_DATE(bar.j,'YYYYMMDD'),'DD-MON-YYYY') AS \"c25\", DECODE(bar.c , 1,'N',0, 'Y', bar.c ) AS \"c26\", bar.y AS y, bar.d, bar.d AS \"c27\" FROM blort.bar , ( SELECT * FROM (SELECT a,a,l,m,b,c, RANK() OVER (PARTITION BY c ORDER BY b DESC) RNK FROM blort.d WHERE y IN (:protocols)) WHERE RNK = 1) qux, blort.foo WHERE bar.c = qux.c(+) AND bar.x = foo.x AND bar.y IN (:protocols) and bar.x IN (:sites)) ) " +
       "SELECT count(*) AS totcount FROM (SELECT \"c1\", \"c2\",\"c3\",\"c4\",\"c5\",\"c6\",\"c7\",\"c8\", \"c9\", \"c10\",\"c11\",\"c12\",\"c13\",\"c14\", \"c15\",\"c16\",\"c17\",\"c18\", \"c19\",\"c20\",\"c21\",\"c22\",\"c23\", \"c24\",\"c25\",\"c26\", \"c27\" FROM (SELECT bar.y AS \"c2\", foo.x AS \"c3\", foo.z AS \"c4\", DECODE(foo.a, NULL,NULL, foo.a ||', '|| foo.b) AS \"c5\" , foo.c AS \"c6\", bar.d AS \"c1\", bar.e AS \"c7\", bar.f AS \"c8\", bar.g AS \"c9\", TO_DATE(TO_CHAR(TO_DATE(bar.h,'YYYYMMDD'),'DD-MON-YYYY'),'DD-MON-YYYY') AS \"c10\", TO_DATE(TO_CHAR(TO_DATE(bar.i,'YYYYMMDD'),'DD-MON-YYYY'),'DD-MON-YYYY') AS \"c11\", CASE WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,'YYYYMMDD'))) > 150 THEN '>150 Days' WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,'YYYYMMDD'))) > 120 THEN '121 to 150 Days' WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,'YYYYMMDD'))) > 90 THEN '91 to 120 Days' WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,'YYYYMMDD'))) > 60 THEN '61 to 90 Days' WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,'YYYYMMDD'))) > 30 THEN '31 to 60 Days' WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,'YYYYMMDD'))) > 0 THEN '1 to 30 Days' ELSE NULL END AS \"c12\", DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,'YYYYMMDD')),NULL) as \"c13\", bar.k AS \"c14\", bar.l ||', '||bar.m AS \"c15\", DECODE(bar.n, NULL, NULL,bar.n ||', '||bar.o) AS \"c16\", bar.p AS \"c17\", bar.q AS \"c18\", bar.r AS \"c19\", bar.s AS \"c20\", qux.a AS \"c21\", TO_CHAR(TO_DATE(qux.b,'YYYYMMDD'),'DD-MON-YYYY') AS \"c22\", DECODE(qux.l,NULL,NULL, qux.l ||', '||qux.m) AS \"c23\", bar.a AS \"c24\", TO_CHAR(TO_DATE(bar.j,'YYYYMMDD'),'DD-MON-YYYY') AS \"c25\", DECODE(bar.c , 1,'N',0, 'Y', bar.c ) AS \"c26\", bar.y AS y, bar.d, bar.d AS \"c27\" FROM blort.bar , ( SELECT * FROM (SELECT a,a,l,m,b,c, RANK() OVER (PARTITION BY c ORDER BY b DESC) RNK FROM blort.d WHERE y IN (:protocols)) WHERE RNK = 1) qux, blort.foo WHERE bar.c = qux.c(+) AND bar.x = foo.x AND bar.y IN (:protocols) and bar.x IN (:sites)) )" | "SELECT count(*) AS totcount FROM (SELECT \"c1\", \"c2\",\"c3\",\"c4\",\"c5\",\"c6\",\"c7\",\"c8\", \"c9\", \"c10\",\"c11\",\"c12\",\"c13\",\"c14\", \"c15\",\"c16\",\"c17\",\"c18\", \"c19\",\"c20\",\"c21\",\"c22\",\"c23\", \"c24\",\"c25\",\"c26\", \"c27\" FROM (SELECT bar.y AS \"c2\", foo.x AS \"c3\", foo.z AS \"c4\", DECODE(foo.a, NULL,NULL, foo.a ||?|| foo.b) AS \"c5\" , foo.c AS \"c6\", bar.d AS \"c1\", bar.e AS \"c7\", bar.f AS \"c8\", bar.g AS \"c9\", TO_DATE(TO_CHAR(TO_DATE(bar.h,?),?),?) AS \"c10\", TO_DATE(TO_CHAR(TO_DATE(bar.i,?),?),?) AS \"c11\", CASE WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,?))) > ? THEN ? WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,?))) > ? THEN ? WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,?))) > ? THEN ? WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,?))) > ? THEN ? WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,?))) > ? THEN ? WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,?))) > ? THEN ? ELSE NULL END AS \"c12\", DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,?)),NULL) as \"c13\", bar.k AS \"c14\", bar.l ||?||bar.m AS \"c15\", DECODE(bar.n, NULL, NULL,bar.n ||?||bar.o) AS \"c16\", bar.p AS \"c17\", bar.q AS \"c18\", bar.r AS \"c19\", bar.s AS \"c20\", qux.a AS \"c21\", TO_CHAR(TO_DATE(qux.b,?),?) AS \"c22\", DECODE(qux.l,NULL,NULL, qux.l ||?||qux.m) AS \"c23\", bar.a AS \"c24\", TO_CHAR(TO_DATE(bar.j,?),?) AS \"c25\", DECODE(bar.c , ?,?,?, ?, bar.c ) AS \"c26\", bar.y AS y, bar.d, bar.d AS \"c27\" FROM blort.bar , ( SELECT * FROM (SELECT a,a,l,m,b,c, RANK() OVER (PARTITION BY c ORDER BY b DESC) RNK FROM blort.d WHERE y IN (:protocols)) WHERE RNK = ?) qux, blort.foo WHERE bar.c = qux.c(+) AND bar.x = foo.x AND bar.y IN (:protocols) and bar.x IN (:sites)) ) " +
       "SELECT count(*) AS totcount FROM (SELECT \"c1\", \"c2\",\"c3\",\"c4\",\"c5\",\"c6\",\"c7\",\"c8\", \"c9\", \"c10\",\"c11\",\"c12\",\"c13\",\"c14\", \"c15\",\"c16\",\"c17\",\"c18\", \"c19\",\"c20\",\"c21\",\"c22\",\"c23\", \"c24\",\"c25\",\"c26\", \"c27\" FROM (SELECT bar.y AS \"c2\", foo.x AS \"c3\", foo.z AS \"c4\", DECODE(foo.a, NULL,NULL, foo.a ||?|| foo.b) AS \"c5\" , foo.c AS \"c6\", bar.d AS \"c1\", bar.e AS \"c7\", bar.f AS \"c8\", bar.g AS \"c9\", TO_DATE(TO_CHAR(TO_DATE(bar.h,?),?),?) AS \"c10\", TO_DATE(TO_CHAR(TO_DATE(bar.i,?),?),?) AS \"c11\", CASE WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,?))) > ? THEN ? WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,?))) > ? THEN ? WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,?))) > ? THEN ? WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,?))) > ? THEN ? WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,?))) > ? THEN ? WHEN DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,?))) > ? THEN ? ELSE NULL END AS \"c12\", DECODE(bar.j, NULL, TRUNC(SYSDATE) - TRUNC(TO_DATE(bar.h,?)),NULL) as \"c13\", bar.k AS \"c14\", bar.l ||?||bar.m AS \"c15\", DECODE(bar.n, NULL, NULL,bar.n ||?||bar.o) AS \"c16\", bar.p AS \"c17\", bar.q AS \"c18\", bar.r AS \"c19\", bar.s AS \"c20\", qux.a AS \"c21\", TO_CHAR(TO_DATE(qux.b,?),?) AS \"c22\", DECODE(qux.l,NULL,NULL, qux.l ||?||qux.m) AS \"c23\", bar.a AS \"c24\", TO_CHAR(TO_DATE(bar.j,?),?) AS \"c25\", DECODE(bar.c , ?,?,?, ?, bar.c ) AS \"c26\", bar.y AS y, bar.d, bar.d AS \"c27\" FROM blort.bar , ( SELECT * FROM (SELECT a,a,l,m,b,c, RANK() OVER (PARTITION BY c ORDER BY b DESC) RNK FROM blort.d WHERE y IN (:protocols)) WHERE RNK = ?) qux, blort.foo WHERE bar.c = qux.c(+) AND bar.x = foo.x AND bar.y IN (:protocols) and bar.x IN (:sites)) )"
