Postgres:通过其索引将JSON数组元素扩展到值

Question

我有一个包含JSON数组的列的表。例如：

with example as (
    select '["a", "b"]'::jsonb as col
    union select ('["c", "d", "e"]'::jsonb) as col
)
select col from example

返回：

col
["a", "b"]
["c", "d", "e"]

我可以使用jsonb_array_elements将每个数组扩展到行：

select jsonb_array_elements(col) from example

返回：

jsonb_array_elements
"a"
"b"
"c"
"d"
"e"

我想要每个数组元素的索引以及元素本身(有点像Python的enumerate(，就像：

jsonb_array_elements    array_index
"a"                     1
"b"                     2
"c"                     1
"d"                     2
"e"                     3

我该怎么做？

我的应用程序只读访问，因此我无法创建函数。

Answer 1

使用 with ordinality：

with example (col) as (
  values 
    ('["a", "b"]'::jsonb),  
    ('["c", "d", "e"]'::jsonb)
)
select t.*
from example, jsonb_array_elements(col) with ordinality as t(e,idx)

返回：

e   | idx
----+----
"a" |   1
"b" |   2
"c" |   1
"d" |   2
"e" |   3

with ordinality仅在使用from子句中的设置返回函数时才能使用，无论如何都强烈推荐。

Answer 2

ahh，这是一个有趣的小邮政难题。下面怎么样？

   WITH example1 AS (
        SELECT '["a", "b"]'::jsonb AS col
    ),
    example2 AS (
        SELECT ('["c", "d", "e"]'::jsonb) AS col
    )
    SELECT  1 AS group, jsonb_array_elements(col) AS jcol, row_number() OVER (ORDER BY jsonb_array_elements(col)) FROM example1
    UNION
    SELECT 2 AS group, jsonb_array_elements(col) AS jcol, row_number() OVER (ORDER BY jsonb_array_elements(col)) FROM example2
    ORDER BY "group", row_number ASC;

会给你

1   "a" 1
1   "b" 2
2   "c" 1
2   "d" 2
2   "e" 3