Which is the fastest hive UDF btw LIKE , get_json_object and json_stupple

Question

I am working on an hive query that needs to count an enormous amount of data reported in a table like this one:

table[ |column_name      | 
       |    (...)        |
       |{"data":"number"}| 
           (...)         ]

The performance of my query is important cause of the number of data I have to treat.

By searching on the web I found three methods to do this. Does some people know which is the fastest? or can propose another method which is faster than those three?

1. SELECT count(*) FROM table WHERE column LIKE TRIM('{"data":"number"}')
   

2. SELECT count(*) FROM table WHERE get_json_object(column,'$.data)=="number"
   

3. SELECT count(*) FROM (
      SELECT json_tuple(column,'number') FROM table) AS new_column
      WHERE new_column.col1 = number