按自定义切片拆分数组

Question

我有一个数组

[('active',1),('active',2),('active',3),('active',4), ('not active',7),('not active',5),('not active',6),('active',7),('active',8),('active',9)]

如果有办法得到这样的数组：

[[('active',1),('active',2),('active',3), ('active',4)],[('not active',7),('not active',5),('not active',6)],[('active',7),('active',8),('active',9)]]

Answer 1

with e as (
    select '1' sku, 'active' status, toInt32('11') dttm
    union all
    select '1' sku, 'active' status, toInt32('12') dttm
    union all
    select '1' sku, 'active' status, toInt32('13') dttm
    union all
    select '1' sku, 'not active' status, toInt32('14') dttm
    union all
    select '1' sku, 'not active' status, toInt32('15') dttm
    union all
    select '1' sku, 'not active' status, toInt32('16') dttm
    union all
    select '1' sku, 'active' status, toInt32('17') dttm
    union all
    select '1' sku, 'active' status, toInt32('18') dttm
    union all
    select '1' sku, 'active' status, toInt32('19') dttm
)
, 'active' as ch
select
    sequence
    , arrayEnumerate(sequence) indexes
    , arraySplit((x, i) -> (sequence.1)[i] = ch and (sequence.1)[i-1] <> ch or (sequence.1)[i] <> ch and (sequence.1)[i - 1] = ch, sequence, indexes) res
from (
    SELECT groupArray(tuple(status, dttm)) as sequence
    from e
    group by sku
) a

输出

+-------------------------------------------------------------------------------------------------------------------------------------------+-------------------------------------------------------------------------------------------------------------------------------------------------+
|sequence                                                                                                                                   |res                                                                                                                                              |
+-------------------------------------------------------------------------------------------------------------------------------------------+-------------------------------------------------------------------------------------------------------------------------------------------------+
|[('active',11),('active',12),('active',13),('not active',14),('not active',15),('not active',16),('active',17),('active',18),('active',19)]|[[('active',11),('active',12),('active',13)],[('not active',14),('not active',15),('not active',16)],[('active',17),('active',18),('active',19)]]|
+-------------------------------------------------------------------------------------------------------------------------------------------+-------------------------------------------------------------------------------------------------------------------------------------------------+

Answer 2

SELECT arraySplit((z, w) -> w, x, arrayMap((i, j) -> ((i.1) != (j.1)), x, arrayPushFront(arrayPopBack(x), ('-', 0)))) AS y
FROM
(
    SELECT [('active', 1), ('active', 2), ('active', 3), ('active', 4), ('not active', 7), ('not active', 5), ('not active', 6), ('active', 7), ('active', 8), ('active', 9)] AS x
)

┌─y─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┐
│ [[('active',1),('active',2),('active',3),('active',4)],[('not active',7),('not active',5),('not active',6)],[('active',7),('active',8),('active',9)]] │
└───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┘

按自定义切片拆分数组

Split array by custom slice

clickhouse