SELECT MAX(col1) FROM<table_name> 子查询，在哪里<table_name>基于来自父查询的值</table_name></table_name>

Question

So lets say we have a one-dimensional table of tables named 'TABLES':

| TABLE_NAMES |
|   'table1'  |
|   'table2'  |
|   'table3'  |
...

Presume each of these tables has an attribute 'quantity' of type INT.

Does there exist a single query that can give me each of these table names, along with a corresponding MAX(quantity) value for that table.

Currently this is my only solution:

SELECT 'table1', MAX(quantity) FROM table1 UNION SELECT 'table2', MAX(quantity) FROM table2...

However, this will not suffice as I do not know what these tables are ahead of time, and will therefore need them inferred from the 'TABLES' table itself.

The below query does not work, however I would imaging the query I am after looks something like this:

SELECT T.TABLE_NAME, (SELECT MAX(quantity) FROM T.TABLE_NAME) FROM TABLES AS T

Appreciate the help,

Thanks

Answer 1

A single query probably not, but a stored procedure probably can provide the output you want.

For example:

create or replace table TABLES_X(tables_names varchar(100));
insert into tables_x values ('table2');
insert into tables_x values ('table3');
insert into tables_x values ('table4');
insert into tables_x values ('table5');

insert into tables_x values ('table2');
insert into tables_x values ('table3');
insert into tables_x values ('table4');
insert into tables_x values ('table5');

insert into tables_x values ('table2');
insert into tables_x values ('table3');
insert into tables_x values ('table4');
insert into tables_x values ('table5');

create or replace table table2 (quantity int);
create or replace table table3 (quantity int);
create or replace table table4 (quantity int);
create or replace table table5 (quantity int);

insert into table2 values (100);
insert into table2 values (200);
insert into table2 values (300); 

insert into table3 values (1000);
insert into table3 values (2000);
insert into table3 values (3000); 

insert into table4 values (10000);
insert into table4 values (20000);
insert into table4 values (30000); 

insert into table5 values (100000);
insert into table5 values (200000);
insert into table5 values (300000); 

Then the stored procedure can look like this:

create or replace procedure read_result_set()
  returns array not null
  language javascript
  as     
  $$
    var array_of_rows = [];
    var my_sql_command = "SELECT DISTINCT(tables_names) FROM TABLES_X";
    var statement1 = snowflake.createStatement( {sqlText: my_sql_command} );
    var result_set1 = statement1.execute();
    
    while (result_set1.next())  {
       var col = result_set1.getColumnValue(1);
       array_of_rows.push(col);
        var sql_cmd = "SELECT MAX(quantity) FROM " + col;
        var statement2 = snowflake.createStatement( {sqlText: sql_cmd} );
        var result_set2 = statement2.execute();
       
        while (result_set2.next())  {
            var col2 = result_set2.getColumnValue(1);       
            array_of_rows.push(col2);
        }
       }
  return array_of_rows
  $$
  ;

Call the procedure:

call read_result_set();

and we get an array that looks like this:

[
          "table2",
          300,
          "table3",
          3000,
          "table4",
          300000,
          "table5",
          300000
]