Remove one, non-unique value from an array

问题

Given this table in PostgreSQL 9.6:

CREATE TABLE test_table (
   id int PRIMARY KEY
 , test_array text[]
);

With a row like:

INSERT INTO test_table (id, test_array)
VALUES (1 , '{A,A,A,B,B,B}');

How would I remove one single 'B' value?

I cannot use:

UPDATE test_table
SET test_array = array_remove(test_array, 'B')
WHERE id = 1;

as it removes all elements of value 'B'. I am just looking to remove a single element (say, the first one).

Any ideas?

回答1:

Based on my old answer on dba.SE that you found and put to good use:

Delete array element by index

You might take it one step further:

CREATE OR REPLACE FUNCTION f_array_remove_elem1(anyarray, anyelement)
  RETURNS anyarray LANGUAGE sql IMMUTABLE AS
'SELECT $1[:idx-1] || $1[idx+1:] FROM array_position($1, $2) idx';

This function takes the value of the element to remove as 2nd parameter. Using the polymorphic pseudo-type anyelement accordingly to make this work for any array type.

Then the UPDATE simply is:

UPDATE test_table
SET    test_array = f_array_remove_elem1(test_array, 'B')
WHERE  id = 1;

db<>fiddle here

While using my original function f_array_remove_elem() that takes the index position instead of the element value, you could do without a subquery:

UPDATE test_table
SET    test_array = f_array_remove_elem(test_array, array_position(test_array, 'B'))
WHERE  id = 1;

Might even be a bit faster than my new function.
And note that the simpler version at the bottom of my old answer works for Postgres 9.6.

回答2:

I found the answer after expanding on the information provided on this page (suggested by Sharon Ben Asher) https://dba.stackexchange.com/questions/94639/delete-array-element-by-index

creating the function mentioned there:

CREATE OR REPLACE FUNCTION f_array_remove_elem(anyarray, int)
RETURNS anyarray LANGUAGE sql IMMUTABLE AS
'SELECT $1[1:$2-1] || $1[$2+1:2147483647]';

and implementing array_position on top of this, the final statement for the example:

UPDATE test_table SET test_array = f_array_remove_elem(test_array, array_position(test_array, 'B')) WHERE id = 1;

回答3:

You can try with this one

select SUBSTRING ( '{A,A,A,B,B,B}',1,position('B' in '{A,A,A,B,B,B}') -1)||SUBSTRING ('{A,A,A,B,B,B}',position('B' in '{A,A,A,B,B,B}')+1)

First find the position of B in the string, then get the substring until just before the first occurrence of B i.e '{A,A,A, then get the substring of the remaining string leaving the B i.e ,B,B}'

来源：https://stackoverflow.com/questions/54992228/remove-one-non-unique-value-from-an-array

标签

sql

arrays

postgresql

sql-update