SSC-EWI-VT0002

Order by table option is not supported in Snowflake

Description

In Vertica, this ORDER BY clause specifies how data is physically sorted within a superprojection, an optimized storage structure for a table. This explicit physical ordering at table creation is not directly supported in Snowflake.

Snowflake handles data storage differently, utilizing micro-partitions. While the data within these micro-partitions can exhibit some natural order based on insertion or if clustering keys are defined, an ORDER BY clause is not used to dictate this physical arrangement during table creation in the same explicit manner as in Vertica's superprojections. Instead, Snowflake employs clustering to optimize data layout for performance, providing a more automated approach to physical ordering.

Code Example

Input Code:

Vertica

IN -> Vertica_01.sql

CREATE TABLE metrics 
(
  metric_id INT,
  business_unit VARCHAR(100),
  metric_category VARCHAR(50) NOT NULL,
  measurement_date DATE NOT NULL
)
ORDER BY measurement_date, business_unit, metric_category;

Output Code:

Snowflake

OUT -> Vertica_01.sql

CREATE TABLE metrics
(
  metric_id INT,
  business_unit VARCHAR(100),
  metric_category VARCHAR(50) NOT NULL,
  measurement_date DATE NOT NULL
)
!!!RESOLVE EWI!!! /*** SSC-EWI-VT0002 - ORDER BY TABLE OPTION IS NOT SUPPORTED IN SNOWFLAKE ***/!!!
ORDER BY measurement_date, business_unit, metric_category
COMMENT = '{ "origin": "sf_sc", "name": "snowconvert", "version": {  "major": 0,  "minor": 0,  "patch": "0" }, "attributes": {  "component": "vertica",  "convertedOn": "06/17/2025",  "domain": "no-domain-provided" }}';

Recommendations

• For Snowflake, the recommendation is to add clustering keys to emulate this behavior, following Snowflake's own recommendations for clustering key implementation.

• If you need more support, you can email us at [e-mail]