如何从postgresql中的数组中排除？

Question

我想在下面的查询中更改数组列call_type的条件，以便排除所有“visit_occurred”的内容。
我如何使用该数组部分？

select staff_id,
   COUNT(event_id) as offered
from call_logs as logs
where visit_offered
and contact_date between now() - interval '1 weeks' and now()
and provider_type = 'Contractor'
and contact_with != 'company_staff'
and direction = 'outbound'
and outcome = 'Successful'
and call_type && array  ['Schedule','schedule_visit']
group by staff_id;

Answer 1

To simply match array elements as a whole use the ARRAY contains operator @>, negated:

AND  NOT call_type @> '{schedule_visit}'::text[]

Note the array wrapper. The cast to text[] is typically not needed, but can help to avoid ambiguity.

If the column can be NULL, and you don't want to exclude such rows:

AND  (call_type @> '{schedule_visit}'::text[]) IS NOT TRUE

Can be supported with an index:

• Can PostgreSQL index array columns?

Pattern matching is more sophisticated. Use a NOT EXISTS expression:

AND NOT EXISTS (
   SELECT FROM unnest(logs.call_type) call
   WHERE  call ILIKE '%visit_occurred%'  -- case insensitive?
   )

No index support for this. Related answer for a JSON array (same principle):

• Search a JSON array for an object containing a value matching a pattern

An alternative would be normalizing your array column as a separate n:1 table.