BigQuery's query filter by json type

Question

My database table like below.

Row	Name	Info
1	Tom	{"count":0, "error":[0,"failed",0]}
2	Joe	{"count":0, "error":["failed","failed","failed"]}
3	Joe	{"count":0, "error":[0,0,0]}
4	Tom	{"count":0, "error":["failed","failed",0]}

I hope filter data by "Info.error's array" in the following order.

1. all is 0
2. exclude all failed
3. at least one is 0

result data

Row	Name	Info
1	Tom	{"count":0, "error":[0,"failed",0]}
3	Joe	{"count":0, "error":[0,0,0]}
4	Tom	{"count":0, "error":["failed","failed",0]}

So, how should enter the SQL syntax? Thanks a lot.

Answer 1

In order to find out the amount of a particular value in a json array you can use the following snippet:

  (SELECT
    COUNTIF(JSON_EXTRACT_SCALAR(error) = "0")
  FROM
    UNNEST(JSON_EXTRACT_ARRAY(ex_table.Info, '$.error')) AS error ) count_zeros,
   
  (SELECT
    COUNTIF(JSON_EXTRACT_SCALAR(error) = "failed")
  FROM
    UNNEST(JSON_EXTRACT_ARRAY(ex_table.Info, '$.error')) AS error ) count_failed

Using these two new variables, you can then create your actual filters.

1.

 WHERE
  count_zero = ARRAY_LENGTH(Info.error)

 WHERE 
  NOT count_failed = ARRAY_LENGTH(Info.error)

WHERE  
  count_zero > 0

Answer 2

You should be able to simplify your rule set as rule 3 accounts for all scenarios you have listed. Given your sample data try the following:

select *,
from sample_data
WHERE "0" in (SELECT * from UNNEST(JSON_VALUE_ARRAY(info, "$.error")))

It produces: