Guard against the sequential traps of Dictionary and SortedDictionary

/*
We know that the traversal order of the Dictionary is the same as that of the Add command, unlike the HashTable command. Therefore, I want to make a decision based on the sequential consistency feature of the Dictionary command, for a Queue with a fixed size of 400, if more than 400 elements are involved, Remove old data and retain new data. Why do I not need Queue because it does not have the ContainsKey search function, because I need to use a key to repeat.
When my program runs for a period of time, a logic problem occurs. Originally, my program logic was 400, and then the old data that was added most was deleted, but the last added data was deleted, when traversal is performed, the order is disordered, and it is not the added program. It is found that the dictionary object cannot be consistent with the program, and the consistency of the Remove sequence cannot be ensured.
*/




SortedDictionary <,> testdict = SortedDictionary <,>, (KeyValuePair <,> kv <,> testdict = Dictionary <,>, (KeyValuePair <,> kv