Question

我是SnowFlake的新手，每天都在探索新事物。我坚持下面的情况。

    SELECT 
   '{COL}' AS field_name,
    a.{COL}AS old_value,
    b.{COL}AS new_value FROM A JOIN B ON(...)
    WHERE a.{COL} != b.{COL}

我想参数化COL。并且它可能具有多个值，例如 COL = col1，col2，col3 。然后我要3个查询，用 UNION 分开，如下所示。

    SELECT 
   'col1' AS field_name,
    a.col1 AS old_value,
    b.col1 AS new_value FROM A JOIN B ON(...)
    WHERE a.col1  != b.col1 

    UNION ALL

    SELECT 
   'col2' AS field_name,
    a.col2 AS old_value,
    b.col2 AS new_value FROM A JOIN B ON(...)
    WHERE a.col2  != b.col2 

    UNION ALL

    SELECT 
   'col3' AS field_name,
    a.col3 AS old_value,
    b.col3 AS new_value FROM A JOIN B ON(...)
    WHERE a.col3  != b.col3

在SnowSQL（SnowFlake）中有什么方法可以实现这一目标？

Answer 1

当然可以在SnowFlake中创建动态SQL。最常见的方法是使用存储过程和/或使用JavaScript的UDF生成SQL并执行它。

在javascript中，您可以使用字符串替换，循环等来创建参数列表，联接条件等。

Here's the general documentation

这是几个有用的javascript片段

//Set Up a multi-column Join Condition based on the columns in ColumnList
//Example output: "a.column1 = b.column1 AND a.column2 = b.column2"
ConditionArray = [];
ColumnList.forEach(function(column){
  conditionArray.push("a." + column+ " = " + "b." + column);
});
joinCondition = conditionArray.join(" AND ");


//Executing a script, and returning the output from a resultSet
//You can chain these calls for cleaner/shorter code
SQLQuery = "SELECT 1;"
SQLStatement = snowflake.createStatement({sqlText: SQLQuery });
SQLResultSet = SQL_Statement.execute();
SQLResultSet.next();
SQLResultSet.getColumnValue(1)

然后，您可以创建循环以执行多个类似的查询，或将它们与UNION组合在一起，然后执行该循环。

https://docs.snowflake.net/manuals/sql-reference/stored-procedures-overview.html

如何以编程方式创建SnowSQL查询

1 个答案: