Introduction

Welcome to Snowflake SnowConvert for Oracle. Let us be your guide on the road to a successful migration.

This is a deprecated version of the SnowConvert documentation, please visit the official site HERE.

What is SnowConvert for Oracle?

SnowConvert is a software that understands Oracle SQL and PL/SQL, and performs the following conversions:

• Oracle SQL to Snowflake SQL

• Oracle PL/SQL to:

  • Snowflake Scripting

  • JavaScript embedded in Snowflake SQL

SnowConvert Terminology

Before we get lost in the magic of these code conversions, here are a few terms/definitions so you know what we mean when we start dropping them all over the documentation:

• SQL (Structured Query Language): The standard language for storing, manipulating, and retrieving data in most modern database architectures.

• PL/SQL: Procedural Language for SQL. This was created by Oracle, and is still used by Oracle as the scripting language for stored procedures and functions in Oracle.

• SnowConvert: The software that converts securely and automatically your Oracle files to the Snowflake cloud data platform.

• Conversion rule or transformation rule: Rules that allow SnowConvert to convert from a portion of source code and determine the expected target code.

• Parse: Parse or parsing is an initial process done by SnowConvert to understand the source code, and build up an internal data structure to process the conversion rules.

Let's dive in to some of the code conversions that Snowflake SnowConvert can perform.

Code Conversions

Oracle SQL to Snowflake SQL

SnowConvert for Oracle takes in Oracle source code in SQL and converts the Data Definition Language (DDL), Data Manipulation Language (DML), and functions in the source code to the corresponding SQL in Snowflake SQL.

Example

Here is an example of the conversion of a simple CREATE TABLE statement.

The source code:

Oracle

CREATE TABLE "MyTable"
(
  "COL1" NUMBER,
  "COL2" NUMBER,
  "COL3" NUMBER GENERATED ALWAYS AS (COL1 * COL2) VIRTUAL,
  "COL4" LONG,
  "COL5" CLOB,
  "COL6" ROWID,
  "COL7" NVARCHAR2(10),
  "COL8" RAW(255),
  CONSTRAINT "PK" PRIMARY KEY ("COL1")
);

The migrated Snowflake SQL code:

Snowflake

CREATE OR REPLACE TABLE PUBLIC."MyTable" (
  "COL1" NUMBER (38,18) /*** MSC-WARNING - MSCEWI1066 - NUMBER TYPE COLUMN MAY NOT BEHAVE SIMILARLY IN SNOWFLAKE ***/,
  "COL2" NUMBER (38,18) /*** MSC-WARNING - MSCEWI1066 - NUMBER TYPE COLUMN MAY NOT BEHAVE SIMILARLY IN SNOWFLAKE ***/,
  /*** MSC-WARNING - MSCEWI3040 - GENERATED ALWAYS STATEMENT REMOVED FROM THE COLUMN DEFINITION ***/
  /*** MSC-WARNING - MSCEWI3040 - VIRTUAL KEYWORD REMOVED FROM THE COLUMN DEFINITION ***/
  "COL3" NUMBER (38,18) /*** MSC-WARNING - MSCEWI1066 - NUMBER TYPE COLUMN MAY NOT BEHAVE SIMILARLY IN SNOWFLAKE ***/ AS (COL1 * COL2),
  "COL4" VARCHAR /*** MSC-WARNING - MSCEWI1036 - LONG DATA TYPE CONVERTED TO VARCHAR ***/,
  "COL5" VARCHAR /*** MSC-WARNING - MSCEWI1036 - CLOB DATA TYPE CONVERTED TO VARCHAR ***/,
  "COL6" VARCHAR(18) /*** MSC-WARNING - MSCEWI1036 - ROWID DATA TYPE CONVERTED TO VARCHAR(18) ***/,
  "COL7" VARCHAR(10),
  "COL8" BINARY /*** MSC-WARNING - MSCEWI1036 - RAW DATA TYPE CONVERTED TO BINARY ***/,
  CONSTRAINT "PK" PRIMARY KEY ("COL1")
);

In this converted SQL you will notice that we are converting many things. A few highlights:

• Adding PUBLIC Schema by default for all the Table and view names if the user doesn't specify one (see how to specify a Schema).
• CREATE TABLE to CREATE OR REPLACE TABLE

• Data Type Conversions:

  • LONG to VARCHAR

  • CLOB to VARCHAR

  • ROWID to VARCHAR

  • NVARCHAR2 to VARCHAR

  • RAW to BINARY

• Data Type Attributes: GENERATED ALWAYS AS (COL1 * COL2) VIRTUAL to AS (COL1 * COL2)

For more information about data types and their equivalent: Data Types. More examples can be found in the rest of the documentation.

Oracle PL/SQL

SnowConvert takes Oracle stored procedures and functions (PL/SQL) and converts them to either Snowflake Scripting or JavaScript embedded into Snowflake SQL. Oracle CREATE PROCEDURE and REPLACE PROCEDURE syntax is replaced by Snowflake CREATE OR REPLACE PROCEDURE syntax.

Example

Here is an example of the conversion of a simple CREATE PROCEDURE in Oracle that does an insert into a table used for logging.

This example will be used for both Snowflake Scripting and JavaScript.

Oracle PL/SQL

CREATE OR REPLACE PROCEDURE SC_DEMO.PROC_LOG 
      (final_proc  VARCHAR2, 
       final_message   VARCHAR2,
       logger_type VARCHAR2 DEFAULT 'I')
AS
BEGIN
  INSERT INTO SC_DEMO.PROC_LOG_TABLE
    VALUES (SC_DEMO.final_logging_seq.NEXTVAL,
            sysdate,
            SUBSTR(logger_type, 1, 1),
            SUBSTR(final_proc, 1, 30),
            SUBSTR(final_message, 1, 1024));
  COMMIT;

END;

To Snowflake Scripting

Snowflake Scripting works as an extension to Snowflake SQL, it adds support for procedural logic and this allow us to create Stored Procedures and replicate similar behaviours and statements of Oracle PL/SQL.

Migrated Example

Snowflake Scripting

CREATE OR REPLACE PROCEDURE SC_DEMO.PROC_LOG
(final_proc VARCHAR, final_message VARCHAR, logger_type VARCHAR
)
RETURNS VARCHAR
LANGUAGE SQL
EXECUTE AS CALLER
ASsql
$$
  BEGIN
    INSERT INTO SC_DEMO.PROC_LOG_TABLE
    VALUES (SC_DEMO.final_logging_seq.NEXTVAL, CURRENT_DATE,
            SUBSTR(:logger_type, 1, 1),
            SUBSTR(:final_proc, 1, 30),
            SUBSTR(:final_message, 1, 1024));
-- ** MSC-ERROR - MSCEWI3111 - COMMIT STATEMENT REQUIRES ADDITIONAL INTERVENTION **
--  COMMIT
        ;
  END;
$$;

To JavaScript

JavaScript is called as a scripting language, all inner statements are converted to JavaScript. If you want to understand better the JavaScript API check this documentation.

Migrated Example

JavaScript in Snow SQL

CREATE OR REPLACE PROCEDURE SC_DEMO.PROC_LOG
(final_proc STRING, final_message STRING, logger_type STRING
)
RETURNS STRING
LANGUAGE JAVASCRIPT
EXECUTE AS CALLER
AS
$$
  // REGION SnowConvert Helpers Code
  // ... Necessary SnowConvert Helpers are inserted here ...
  // END REGION

  EXEC(`INSERT INTO SC_DEMO.PROC_LOG_TABLE
    VALUES (SC_DEMO.final_logging_seq.NEXTVAL, CURRENT_DATE,
            SUBSTR(?, 1, 1),
            SUBSTR(?, 1, 30),
            SUBSTR(?, 1, 1024))`,[LOGGER_TYPE,FINAL_PROC,FINAL_MESSAGE]);
  EXEC(`-- ** MSC-ERROR - MSCEWI3111 - COMMIT STATEMENT REQUIRES ADDITIONAL INTERVENTION **
--  COMMIT
        ;`);
$$;

In this converted SQL you will notice that we have converted to a new language (JavaScript) embedded into Snowflake SQL. There are more than a few highlights, but suffice it to say that this documentation has all the essentials to understand this kind of conversion.

The line that states // ... Necessary SnowConvert Helpers are inserted here ... will actually have the SnowConvert JavaScript Helpers. They can be lengthy, so they are removed from this first example.

And that's it! Snowflake SnowConvert takes the pain and frustration out of changing data platforms. Learn more about getting started with SnowConvert for Oracle on the next page.