如何根据 postgresql 中数组列中的值将行拆分为多行？ （参见说明中的示例）

Question

ia have postgresql query that gets results like this:

id | arrayElements
 0 | [0, 2, 1]
 1 | [0, 3]
 2 | [1]

In the above example the values inside arrayElements means something. I want to split this table based on those arrayElements (to get count).

I want to convert it to:

id | arrayElement
 0 | 0
 0 | 2
 0 | 1
 1 | 0
 1 | 3
 2 | 1

here i'm essentially spliting rows by elements inside arrayElements column. arrayElements column contains string data.

i am using postgresql 14.

my end goal is to get the count per arrayElement which is why i'm separating them into different rows.

how should I solve this?

Answer 1

Also, if you want to get count of arrayElements , you can do this:

select sum(cardinality(arrayElements)) from tableName;

Answer 2

select id, UNNEST(string_to_array(arrayElements, ',') as arrayElement;

should do it. In my case arrayElement was a string that I had to convert to an array.

There may be the case that elements might have a '[' or ']' at the beginning or end so to fix that you can modify the query to use SUBSTRING and remove those square brackets:

SELECT
  SUBSTRING(UNNEST(string_to_array(arrayElements, ',')), '([0-9]{1,10})') AS "arrayElement",
  COUNT(id)
FROM table
GROUP BY arrayElement

the regex used in substring assumes that the maximum number of digits arrayElement can have is 10.