Can't detect null value from JSON_EXTRACT - mysql

Question

I have a database that has an array of data stored in a JSON column. I need to find all values that have a null value at a particular position in the JSON array. While pulling out the data with JSON_EXTRACT seemed trivial, none of my comparisons to null have worked, all of them claiming the value is null.
Here is the example code that should work as far as I can tell:
SELECT JSON_EXTRACT(`COLUMNS_HEADERS`, '$[1]') , (JSON_EXTRACT(`COLUMNS_HEADERS`, '$[1]') is null)
FROM ate.readings_columns_new;
The first few rows of my results table look like this:
null | 0
"INTERNALTEMPERATURE" | 0
"INPUT_VOLTAGE" | 0
null | 0
null | 0
"AH1" | 0
I have tried every comparison I can think of, and they all result in a 0:
(JSON_EXTRACT(`COLUMNS_HEADERS`, '$[1]') is null)
(JSON_EXTRACT(`COLUMNS_HEADERS`, '$[1]') <=> null)
ISNULL(JSON_EXTRACT(`COLUMNS_HEADERS`, '$[1]'))
(JSON_EXTRACT(`COLUMNS_HEADERS`, '$[1]') <=> 'null')
Is there some key to comparing null values pulled from a JSON_EXTRACT?

Answer 1

SELECT
JSON_EXTRACT(`COLUMNS_HEADERS`, '$[1]'),
(JSON_EXTRACT(`COLUMNS_HEADERS`, '$[1]') = CAST('null' AS JSON))
FROM ate.readings_columns_new;
or
SELECT
JSON_EXTRACT(`COLUMNS_HEADERS`, '$[1]'),
(JSON_TYPE(JSON_EXTRACT(`COLUMNS_HEADERS`, '$[1]')) = 'NULL')
FROM ate.readings_columns_new;
See the docs for JSON_TYPE.

Answer 2

A bit of a belated answer but I just hit this problem and couldn't find anything reasonably documented. The solution I ended ended up using was the json_type function as 'abl' pointed out above.
The trick was to compare with the string 'NULL' not null or NULL.
As a test throw the following into a mysql prompt and play around with the values
(if using phpMyAdmin don't forget to check 'show this query here again' and 'retain query box' - the universe is frustrating enough without losing edits..)
set #a='{"a":3,"b":null}';
select if(json_type(json_extract(#a,'$.b')) = 'NULL',1,0);
I ended up with the following.
mysql> set #a='{"a":3,"b":null}';
Query OK, 0 rows affected (0.00 sec)
mysql> select if(json_type(json_extract(#a,'$.b')) = 'NULL',1,0);
+----------------------------------------------------+
| if(json_type(json_extract(#a,'$.b')) = 'NULL',1,0) |
+----------------------------------------------------+
| 1 |
+----------------------------------------------------+
1 row in set (0.00 sec)
mysql> set #a='{"a":3,"b":1}';
Query OK, 0 rows affected (0.00 sec)
mysql> select if(json_type(json_extract(#a,'$.b')) = 'NULL',1,0);
+----------------------------------------------------+
| if(json_type(json_extract(#a,'$.b')) = 'NULL',1,0) |
+----------------------------------------------------+
| 0 |
+----------------------------------------------------+
1 row in set (0.00 sec)
As the bare bones of a stored procedure - which is what I needed it for - using the 'if' statements rather than the if() function.
drop procedure if exists test;
delimiter $$
create procedure test(in x json)
begin
if json_type(json_extract(x,'$.b')) = 'NULL' then
select 1;
else
select 0;
end if;
end$$
delimiter;
mysql> call test('{"a":3,"b":1}');
+---+
| 0 |
+---+
| 0 |
+---+
1 row in set (0.00 sec)
Query OK, 0 rows affected (0.00 sec)
mysql> call test('{"a":3,"b":null}');
+---+
| 1 |
+---+
| 1 |
+---+
1 row in set (0.00 sec)
Query OK, 0 rows affected (0.00 sec)

Answer 3

Well I had a suspicion but I found a workaround that confirms that a JSON null value is not the same as a MySQL null value.
I tried various methods to get a similar null value but the only one that works is to extract a null JSON value from an array like the value I'm attempting to check against:
SELECT JSON_EXTRACT(`COLUMNS_HEADERS`, '$[1]') , (JSON_EXTRACT(`COLUMNS_HEADERS`, '$[1]') = JSON_EXTRACT('[null]', '$[0]'))
FROM ate.readings_columns_new;
This seems like bad form, but was the only way I could get a value that evaluated as equal to the null values in my array.

Answer 4

Another trick is MySQL's NULLIF function
SELECT COLUMNS_HEADERS->>"$[1]", NULLIF(COLUMNS_HEADERS->>"$[1]",'null') IS NULL)
(I'm also using ->> which is an alias for JSON_UNQUOTE(JSON_EXTRACT())
That way querying a column containing {"id":1},{"id":2},{"id":null} & {"name":4} for the JSON path $.id will return 1,2,NULL,NULL instead of 1,2,null,NULL