TOMBSTONE Messages : (ksqlDB Tutorial) How to find distinct values in a stream of events

For the tutorial below
https://kafka-tutorials.confluent.io/finding-distinct-events/ksql.html

If we don't make use of "TIMESTAMP", I see TOMBSTONE Messages. What is the reason for this?

The CLICKS stream is modified to not include the TIMESTAMPA table is created as below
```
CREATE STREAM CLICKS (IP_ADDRESS STRING, URL STRING)
WITH (KAFKA_TOPIC = 'CLICKS',
FORMAT = 'JSON',
PARTITIONS = 1);
```

We then insert the values
```
INSERT INTO CLICKS (IP_ADDRESS, URL) VALUES ('10.0.0.1', 'https://docs.confluent.io/current/tutorials/examples/kubernetes/gke-base/docs/index.html');
INSERT INTO CLICKS (IP_ADDRESS, URL) VALUES ('10.0.0.12', 'https://www.confluent.io/hub/confluentinc/kafka-connect-datagen');
INSERT INTO CLICKS (IP_ADDRESS, URL) VALUES ('10.0.0.13', 'https://www.confluent.io/hub/confluentinc/kafka-connect-datagen');

INSERT INTO CLICKS (IP_ADDRESS, URL) VALUES ('10.0.0.1', 'https://docs.confluent.io/current/tutorials/examples/kubernetes/gke-base/docs/index.html');
INSERT INTO CLICKS (IP_ADDRESS, URL) VALUES ('10.0.0.12', 'https://www.confluent.io/hub/confluentinc/kafka-connect-datagen');
INSERT INTO CLICKS (IP_ADDRESS, URL) VALUES ('10.0.0.13', 'https://www.confluent.io/hub/confluentinc/kafka-connect-datagen');
```

We can then query the STREAM to see duplicate clicks
```
ksql> SELECT
IP_ADDRESS,
URL
FROM CLICKS
GROUP BY IP_ADDRESS, URL
HAVING COUNT(IP_ADDRESS) = 1
EMIT CHANGES;

+----------------------------------------------------------------+----------------------------------------------------------------+
|IP_ADDRESS |URL |
+----------------------------------------------------------------+----------------------------------------------------------------+
|10.0.0.1 |https://docs.confluent.io/current/tutorials/examples/kubernetes/|
| |gke-base/docs/index.html |
|10.0.0.12 |https://www.confluent.io/hub/confluentinc/kafka-connect-datagen |
|10.0.0.13 |https://www.confluent.io/hub/confluentinc/kafka-connect-datagen |
|10.0.0.1 |https://docs.confluent.io/current/tutorials/examples/kubernetes/|
| |gke-base/docs/index.html |
|10.0.0.12 |https://www.confluent.io/hub/confluentinc/kafka-connect-datagen |
|10.0.0.13 |https://www.confluent.io/hub/confluentinc/kafka-connect-datagen |

```
We can also create the table as per the lesson and query on the table to find the TOMBSTONE messages.
```
CREATE TABLE DETECTED_CLICKS AS
SELECT
IP_ADDRESS AS KEY1,
URL AS KEY2,
AS_VALUE(IP_ADDRESS) AS IP_ADDRESS,
AS_VALUE(URL) AS URL
FROM CLICKS
GROUP BY IP_ADDRESS, URL
HAVING COUNT(IP_ADDRESS) = 1;

SELECT * FROM DETECTED_CLICKS EMIT CHANGES;

+-------------------------------+-------------------------------+-------------------------------+-------------------------------+
|KEY1 |KEY2 |IP_ADDRESS |URL |
+-------------------------------+-------------------------------+-------------------------------+-------------------------------+
|10.0.0.1 |https://docs.confluent.io/curre|10.0.0.1 |https://docs.confluent.io/curre|
| |nt/tutorials/examples/kubernete| |nt/tutorials/examples/kubernete|
| |s/gke-base/docs/index.html | |s/gke-base/docs/index.html |
|10.0.0.12 |https://www.confluent.io/hub/co|10.0.0.12 |https://www.confluent.io/hub/co|
| |nfluentinc/kafka-connect-datage| |nfluentinc/kafka-connect-datage|
| |n | |n |
|10.0.0.13 |https://www.confluent.io/hub/co|10.0.0.13 |https://www.confluent.io/hub/co|
| |nfluentinc/kafka-connect-datage| |nfluentinc/kafka-connect-datage|
| |n | |n |
|10.0.0.1 |https://docs.confluent.io/curre| | |
| |nt/tutorials/examples/kubernete| | |
| |s/gke-base/docs/index.html | | |
|10.0.0.12 |https://www.confluent.io/hub/co| | |
| |nfluentinc/kafka-connect-datage| | |
| |n | | |
|10.0.0.13 |https://www.confluent.io/hub/co| | |
| |nfluentinc/kafka-connect-datage| | |
| |n | | |
```

What is the magic behind the TIMESTAMP field that leads to a total different behavior that we see in the lesson explanation?

Much appreciated.