dataroaring commented on code in PR #13056:
URL: https://github.com/apache/doris/pull/13056#discussion_r989830599


##########
be/src/exprs/aggregate_functions.cpp:
##########
@@ -2534,6 +2535,97 @@ IntVal 
AggregateFunctions::window_funnel_finalize(FunctionContext* ctx, const St
     return doris_udf::IntVal(val);
 }
 
+// Refer to AggregateFunctionRetention.h in 
https://github.com/ClickHouse/ClickHouse.git
+struct RetentionState {
+    static constexpr size_t max_events = 32;
+    std::bitset<max_events> events;
+
+    RetentionState() {}
+
+    void add(int event) { events.set(event); }
+
+    void merge(RetentionState* other) { events |= other->events; }
+
+    BooleanVal* getRetentionData() {
+        static BooleanVal data_to[max_events];
+        int current_offset = 0;
+
+        const bool first_flag = events.test(0);
+        data_to[current_offset] = first_flag;
+        ++current_offset;
+
+        for (size_t i = 1; i < events.size(); ++i) {
+            data_to[current_offset] = (first_flag && events.test(i));
+            ++current_offset;
+        }
+        return data_to;
+    }
+
+    int64_t serizlized_size() { return sizeof(events); }
+
+    void serialize(uint8_t* buf) {
+        memcpy(buf, &events, sizeof(events));

Review Comment:
   bitset is a container, memcpy should not be used for serialized. Serialized 
is used to exchange infomation across processes. e.g. serialized data is 
transfered from doris_be in machine A to doris_be in machine B. So we can not 
use memory structure directly.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscr...@doris.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@doris.apache.org
For additional commands, e-mail: commits-h...@doris.apache.org

Reply via email to