executor: convert json numeric value to float64 in hash (#38065)

close pingcap/tidb#38049

executor/aggfuncs/aggfunc_test.go

@@ -189,8 +189,7 @@ func distinctUpdateMemDeltaGens(srcChk *chunk.Chunk, dataType *types.FieldType)
 		case mysql.TypeJSON:
 			jsonVal := row.GetJSON(0)
 			bytes := make([]byte, 0)
-			bytes = append(bytes, jsonVal.TypeCode)
-			bytes = append(bytes, jsonVal.Value...)
+			bytes = jsonVal.HashValue(bytes)
 			val = string(bytes)
 			memDelta = int64(len(val))
 		default:

executor/aggfuncs/func_count_distinct.go

@@ -402,7 +402,7 @@ func evalAndEncode(
 		if err != nil || isNull {
 			break
 		}
-		encodedBytes = appendJSON(encodedBytes, buf, val)
+		encodedBytes = val.HashValue(encodedBytes)
 	case types.ETString:
 		var val string
 		val, isNull, err = arg.EvalString(sctx, row)

executor/executor_test.go

@@ -5981,6 +5981,40 @@ func TestIsFastPlan(t *testing.T) {
 	}
 }
 
+func TestCountDistinctJSON(t *testing.T) {
+	store := testkit.CreateMockStore(t)
+
+	tk := testkit.NewTestKit(t, store)
+	tk.MustExec("use test")
+
+	tk.MustExec("drop table if exists t")
+	tk.MustExec("create table t(j JSON)")
+	tk.MustExec("insert into t values('2010')")
+	tk.MustExec("insert into t values('2011')")
+	tk.MustExec("insert into t values('2012')")
+	tk.MustExec("insert into t values('2010.000')")
+	tk.MustExec("insert into t values(cast(? as JSON))", uint64(math.MaxUint64))
+	tk.MustExec("insert into t values(cast(? as JSON))", float64(math.MaxUint64))
+
+	tk.MustQuery("select count(distinct j) from t").Check(testkit.Rows("5"))
+}
+
+func TestHashJoinJSON(t *testing.T) {
+	store := testkit.CreateMockStore(t)
+
+	tk := testkit.NewTestKit(t, store)
+	tk.MustExec("use test")
+
+	tk.MustExec("drop table if exists t")
+	tk.MustExec("create table t(id int(11), j JSON, d DOUBLE)")
+	tk.MustExec("insert into t values(0, '2010', 2010)")
+	tk.MustExec("insert into t values(1, '2011', 2011)")
+	tk.MustExec("insert into t values(2, '2012', 2012)")
+	tk.MustExec("insert into t values(3, cast(? as JSON), ?)", uint64(math.MaxUint64), float64(math.MaxUint64))
+
+	tk.MustQuery("select /*+inl_hash_join(t2)*/ t1.id, t2.id from t t1 join t t2 on t1.j = t2.d;").Check(testkit.Rows("0 0", "1 1", "2 2"))
+}
+
 func TestBinaryStrNumericOperator(t *testing.T) {
 	store := testkit.CreateMockStore(t)
 

types/json_binary.go

@@ -538,6 +538,12 @@ func (bj BinaryJSON) HashValue(buf []byte) []byte {
 		} else {
 			buf = append(buf, bj.Value...)
 		}
+	case JSONTypeCodeUint64:
+		if bj.GetUint64() == uint64(float64(bj.GetUint64())) {
+			buf = appendBinaryFloat64(buf, float64(bj.GetUint64()))
+		} else {
+			buf = append(buf, bj.Value...)
+		}
 	case JSONTypeCodeArray:
 		elemCount := int(jsonEndian.Uint32(bj.Value))
 		for i := 0; i < elemCount; i++ {

util/codec/codec.go

@@ -386,7 +386,8 @@ func encodeHashChunkRowIdx(sc *stmtctx.StatementContext, row chunk.Row, tp *type
 		b = (*[unsafe.Sizeof(v)]byte)(unsafe.Pointer(&v))[:]
 	case mysql.TypeJSON:
 		flag = jsonFlag
-		b = row.GetBytes(idx)
+		json := row.GetJSON(idx)
+		b = json.HashValue(b)
 	default:
 		return 0, nil, errors.Errorf("unsupport column type for encode %d", tp.GetType())
 	}
@@ -645,7 +646,9 @@ func HashChunkSelected(sc *stmtctx.StatementContext, h []hash.Hash64, chk *chunk
 				isNull[i] = !ignoreNull
 			} else {
 				buf[0] = jsonFlag
-				b = column.GetBytes(i)
+				json := column.GetJSON(i)
+				b = b[:0]
+				b = json.HashValue(b)
 			}
 
 			// As the golang doc described, `Hash.Write` never returns an error..

util/codec/codec_test.go

@@ -1187,6 +1187,9 @@ func TestHashChunkRow(t *testing.T) {
 
 	testHashChunkRowEqual(t, "x", []byte("x"), true)
 	testHashChunkRowEqual(t, "x", []byte("y"), false)
+
+	testHashChunkRowEqual(t, types.CreateBinaryJSON(int64(1)), types.CreateBinaryJSON(float64(1.0)), true)
+	testHashChunkRowEqual(t, types.CreateBinaryJSON(uint64(math.MaxUint64)), types.CreateBinaryJSON(float64(math.MaxUint64)), false)
 }
 
 func TestValueSizeOfSignedInt(t *testing.T) {