Record ranking in Clickhouse - clickhouse

Question

For example, i have a table:
CREATE DATABASE IF NOT EXISTS example;
CREATE TABLE IF NOT EXISTS example.etable (starttime datetime, name string,) ENGINE = MergeTree;
In order to apply the GROUP BY operation , I need to calculate the rank for each record, so that consecutive ordered records with the same value of the "name" field have the same rank. If the current record has a value of the name field other than the previous one, then the rank is incremented.
In MySQL this can be done with query like that:
SELECT name, starttime,
#prev := #curr,
#curr := name,
#rank := IF(#prev = #curr, #rank, #rank+1) AS rank
FROM example.etable,
(SELECT #curr := null, #prev := null, #rank := 0) r
ORDER BY starttime ASC;
Example output:
+------+---------------------+----------------+---------------+------+
| name | starttime | #prev := #curr | #curr := name | rank |
+------+---------------------+----------------+---------------+------+
| s1 | 2020-05-14 15:56:46 | NULL | s1 | 1 |
| s1 | 2020-05-14 15:56:49 | s1 | s1 | 1 |
| s1 | 2020-05-14 15:56:51 | s1 | s1 | 1 |
| s2 | 2020-05-14 15:56:53 | s1 | s2 | 2 |
| s1 | 2020-05-14 15:56:56 | s2 | s1 | 3 |
| s3 | 2020-05-14 15:56:59 | s1 | s3 | 4 |
+------+---------------------+----------------+---------------+------+
So, here is the question, how can I achieve this in Clickhouse?

Answer 1

Calculation rank consist of three steps:
transform relation to arrays (groupArray)
calculate ranks (arrayCumSum)
transform arrays to relation (arrayJoin).
SELECT result.1 starttime, result.2 name, result.3 rank
FROM (
SELECT
groupArray(starttime) starttime_arr,
groupArray(name) name_arr,
arrayCumSum((name, index) -> index = 1 ? 1 : (name_arr[index - 1] = name ? 0 : 1), name_arr, arrayEnumerate(name_arr)) ranks,
arrayZip(starttime_arr, name_arr, ranks) result_array,
arrayJoin(result_array) result
FROM (
SELECT *
FROM (
/* emulate the 'example.etable'-table */
SELECT toDateTime(test_data.1) AS starttime, test_data.2 AS name
FROM (
SELECT arrayJoin([
('2020-05-14 15:56:46', 's1'),
('2020-05-14 15:56:49', 's1'),
('2020-05-14 15:56:51', 's1'),
('2020-05-14 15:56:53', 's2'),
('2020-05-14 15:56:56', 's1'),
('2020-05-14 15:56:59', 's3')
]) test_data))
ORDER BY starttime)
)
/* result
┌───────────starttime─┬─name─┬─rank─┐
│ 2020-05-14 15:56:46 │ s1 │ 1 │
│ 2020-05-14 15:56:49 │ s1 │ 1 │
│ 2020-05-14 15:56:51 │ s1 │ 1 │
│ 2020-05-14 15:56:53 │ s2 │ 2 │
│ 2020-05-14 15:56:56 │ s1 │ 3 │
│ 2020-05-14 15:56:59 │ s3 │ 4 │
└─────────────────────┴──────┴──────┘
*/