SnowConvert AI - Top-Level Code Units Report¶

什么是顶级代码单元？¶

顾名思义，代码单元是最基础、可独立执行的最小元素。在大多数情况下，这些单元是语句，但它们也包括脚本文件，因为它们是作为单个元素执行。

一些代码单元可以嵌套在其他代码单元中。在代码单元层级结构中，当某个代码单元上方不存在其他上级单元时，该单元即被称为顶级代码单元

什么是超出范围的顶级代码单元？¶

Out-of-scope Top-Level Code Units are Code Units that are out of the conversion scope of SnowConvert AI. Because of this, these code units are not considered when calculating the conversion rate. Each of these code units will not have a conversion rate (it will appear as N/A).

如果输入代码仅包含超出范围的代码单元，则整个迁移的代码行转换率将为 0%。

以下 CREATE TRIGGER 被视为超出范围的代码单元。

CREATE OR REPLACE TRIGGER my_trigger
    AFTER 
    UPDATE
    ON my_table
    FOR EACH ROW
BEGIN
   NULL;
END;

顶级代码单元示例¶

在下一部分中，我们可以看到一些顶级代码单元的示例。

SELECT * FROM table1;

SELECT * FROM (SELECT * FROM table1);

CREATE VIEW view1 AS SELECT * FROM table1;

CREATE PROCEDURE procedure1
AS
BEGIN
    DELETE FROM table1;
END;

COMMIT;

CREATE PACKAGE package_body1 IS
    FUNCTION function1
    RETURN VARCHAR
    IS
    BEGIN
        RETURN 'HELLO'';
    END;
END;

.LOGON e/fml,notebook
.COMPILE FILE = example.spl;
COMMIT;
CALL samplesp1 (8888, pAmount);
.LOGOFF

DECLARE @Counter int;  
SET @Counter = 1;  
WHILE @Counter < 10  
BEGIN   
    SELECT @Counter  
    SET @Counter = @Counter + 1  
    IF @Counter = 4 GOTO Branch_One
    IF @Counter = 5 GOTO Branch_Two  
END  
Branch_One:  
    SELECT 'Jumping To Branch One.'  
    GOTO Branch_Three;
Branch_Two:  
    SELECT 'Jumping To Branch Two.'  
Branch_Three:  
    SELECT 'Jumping To Branch Three.';
GO

在其他报告中，代码单元方法论是如何体现的？¶

其他报告中也介绍了代码单元方法论。本部分说明这些值如何显示，或者如何与其他报告相关。

顶级代码单元报告中的信息¶

示例¶

假设 ORACLE SQL 中的以下 CREATE TABLE 位于其名为 table_example.sql 的文件中。

CREATE TABLE my_table (
  my_column DATE DEFAULT TO_DATE(CURRENT_DATE, 'J'),
  NOT A VALID COLUMN
);

CREATE OR REPLACE TABLE my_table (
   my_column TIMESTAMP /*** SSC-FDM-OR0042 - DATE TYPE COLUMN HAS A DIFFERENT BEHAVIOR IN SNOWFLAKE. ***/ DEFAULT PUBLIC.JULIAN_TO_GREGORIAN_DATE_UDF(CURRENT_DATE(), 'J')
--                                                                                                                                                                          ,
-- ** SSC-EWI-0001 - UNRECOGNIZED TOKEN ON LINE '3' COLUMN '3' OF THE SOURCE CODE STARTING AT 'NOT'. EXPECTED 'Column Definition' GRAMMAR. LAST MATCHING TOKEN WAS ',' ON LINE '2' COLUMN '52'. FAILED TOKEN WAS 'NOT' ON LINE '3' COLUMN '3'. CODE '15'. **
--  NOT A VALID COLUMN
 )
 COMMENT = '{"origin":"sf_sc","name":"snowconvert","version":{"major":1, "minor":0},"attributes":{"component":"oracle"}}'
 ;

顶级代码单元报告将包含前面显示的表的单个条目。

以下是将在此 CREATE TABLE 语句的条目中报告的所有值：

• 分区键 值将取决于迁移，因此此处的值会有所不同。

• 文件类型 将是 SQL，因为它是在扩展名为 .sql 的文件上迁移的。

• 类别 将是 TABLE，因为 CREATE TABLE 语句是 TABLE 代码单元类别的一部分。

• 代码单元 本身将是 CREATE TABLE。

• 找到此代码单元的 文件名 将是 table_example.sql。

• 假设 CREATE TABLE 语句位于文件的开头，则 行号 将为 1。

• 代码行数 的数字将为 4。

• EWI 计数 列将报告 1，因为输出代码有一个解析 EWI。

• FDM 计数 列将报告 1，因为输出代码存在与数据类型相关的 FDM 问题。

• PRF 计数 将报告 0，因为输出代码中不存在 PRF 问题。

• 在这种情况下，最高 EWI 严重性 将为“严重”，因为这是示例解析 EWI 的严重性。另一个的严重性为“低”。

• 使用的 UDFs 列将是 JULIAN_TO_GREGORIAN_DATE_UDF，因为此自定义的用户定义函数已添加，以转换输入代码的 TO_DATE 函数。

• EWI 列将报告“SSC-EWI-0001”，因为这是输出代码中添加的 EWIs 之一。

• ** FDM ** 列将报告“SSC-FDM-OR0042”，因为这是输出代码中添加的 FDMs 之一。

• PRF 列将报告“不适用”，因为输出代码中不存在 PRF 问题。

• 转换状态 将为“部分”，因为如果没有 EWIs，该代码单元中只有部分片段能够迁移。

• LoC 转换百分比 为 50%，因为在 4 行中，只有 2 行成功转换。

Deployment Order¶

The deployment order column represents the correct order to deploy each code unit into Snowflake.

The following code exemplifies in depth how the deployment order is calculated.

CREATE TABLE TABLE1 ( -- level 0, no dependencies
   COL1 INT
);

CREATE TABLE TABLE2 ( -- level 0, no dependencies
   COL1 INT
);

CREATE VIEW VIEW1 -- level 4, depends on level-3 objects
AS SELECT * FROM VIEW2, VIEW3;

CREATE VIEW VIEW2 -- level 3, depends on level-2 objects
AS SELECT * FROM VIEW4, VIEW5, VIEW3;

CREATE VIEW VIEW4 -- level 1, depends on level-0 objects
AS SELECT * FROM TABLE1, TABLE2;

CREATE VIEW VIEW5 -- level 1, depends on level-0 objects
AS SELECT * FROM TABLE1;

CREATE VIEW VIEW3 -- level 2, depends on level-1 objects
AS SELECT * FROM VIEW6;

CREATE VIEW VIEW6 -- level 1, depends on level-0 objects
AS SELECT * FROM TABLE2;

The deployment order starts with 0, so code units without any dependencies will start at this level. In the example above, TABLE1 and TABLE2 will have a level 0 .

For the next level, we will focus on code units that depend on code units of level 0. VIEW4, VIEW5, and VIEW6 depend directly on TABLE1 and TABLE2, so their level will be 1.

After identifying all the code units of level 1 , we will focus on code units of level 2. In that particular scenario, just VIEW3 depends on VIEW6 , so VIEW3 will be level 2.

Once we identify all code units of level 2, we will focus on level 3. In the example above, VIEW2 depends on VIEW4, VIEW5 and VIEW3, however, the highest dependency level is 2, so, VIEW2 will be of level 3.

Finally, we got VIEW1, which depends on VIEW2 and VIEW3. Since VIEW2 is the dependency with higher level, VIEW1 will get level 4.

After making all the calculations, the top-level code units report will look something like the following table.

CREATE TABLE TABLE1 ( -- level 0, no dependencies
  COL1 INT
);

CREATE VIEW VIEW1 -- level 1*, depends on level-0 objects and has a missing dependency
AS SELECT * FROM TABLE1, TABLE2;

CREATE VIEW VIEW2 -- level 2*, depends on level-1* objects 
AS SELECT * FROM VIEW1;

CREATE DATABASE LINK DBLINK1
CONNECT TO PUBLIC IDENTIFIED BY VALUES ':1'
USING 'TEST';

CREATE MATERIALIZED VIEW VIEW1 REFRESH WITH ROWID 
AS SELECT * FROM TABLE1@DBLINK1;

CREATE TABLE TABLE1 (
  COL1 INT
);

CREATE OR REPLACE PROCEDURE PROC1 (param1 NUMBER)
IS
BEGIN
    CREATE VIEW VIEW1 
    AS
    SELECT * FROM TABLE1;
END;

CREATE VIEW VIEW2
AS SELECT * FROM VIEW1;