Author Topic: Incomplete or Buggy Snowflake DB Engineering  (Read 1205 times)

sousac

  • EA User
  • **
  • Posts: 21
  • Karma: +2/-0
    • View Profile
    • Integrationworx, a Sparx EA Authorized Training Partner
Incomplete or Buggy Snowflake DB Engineering
« on: July 11, 2025, 01:58:17 pm »
We are using v 17.1 which offers support for DW platforms including Snowflake.

We are seeing catastrophic errors/omissions across our workstations (not workstation/installation specific) and I think they are bugs.

When working with a model that was originally authored/created in V16, we see the following behaviour.

#1: No DDL Template content for Snowflake
- 17.1 appears to have empty (no template content) for DDL generation. The basic create table is still generated despite having no actual template.  Copying over general templates for table definition and FK declaration (and subordinate templates) does not result in any FK declarations being added to the DDL.

With a brand new model, created fresh in v 17.1, we still see issues with data type mapping.

#2: Datatype Mapping / Conversion Missing for Snowflake, Cannot Save Common Types
- If a table is set to SQL Server 2012, for example, changing it to Snowflake should allow for datatype conversion/mapping. Instead Sparx wipes out all the data type information when this is done.
- Attempts to declare a common data type (on the data type dialogue) result in an error indicating "Technology data type cannot be saved". A common type is required to make entries in the data type mapping.
- The data type mapping, as a result, only allows mapping for Snowflake datatypes that I fictionally create (none of the built in ones can be chosen because the don't have a common type).



« Last Edit: July 11, 2025, 02:07:27 pm by sousac »
- Claudio

Richard Freggi

  • EA User
  • **
  • Posts: 493
  • Karma: +18/-7
    • View Profile
Re: Incomplete or Buggy Snowflake DB Engineering
« Reply #1 on: July 11, 2025, 10:05:38 pm »
I also have issues with Snowflake in EA 17.1.  I updated my license from 16 after obtaining written confirmation from Sparx Sales that 17.1 would reverse engineer Snowflake DDL.  it sorta kinda does -ish as long as the DDL is very plain.  In my case reverse engineering fails when views DDL has column masking statements.  So I have to manually edit the DDL before it's run thru EA. 

I really expect that Sparx matures its Snowflake offering before my license needs renewal!

Modesto Vega

  • EA Practitioner
  • ***
  • Posts: 1137
  • Karma: +30/-8
    • View Profile
Re: Incomplete or Buggy Snowflake DB Engineering
« Reply #2 on: July 11, 2025, 10:13:25 pm »
I would expect that of a 1st release. Snowflake is not a Microsoft product, it is an Azure Marketplace offering in some degree of partnering with Microsoft.

#2: Datatype Mapping / Conversion Missing for Snowflake, Cannot Save Common Types
- If a table is set to SQL Server 2012, for example, changing it to Snowflake should allow for datatype conversion/mapping. Instead Sparx wipes out all the data type information when this is done.

That is very easy to say and not so easy to implement. AFIK, Sparx EA does not support that for any language.

Richard Freggi

  • EA User
  • **
  • Posts: 493
  • Karma: +18/-7
    • View Profile
Re: Incomplete or Buggy Snowflake DB Engineering
« Reply #3 on: July 19, 2025, 12:48:57 pm »
I had a few emails with Sparx sales and support because I bought the 17.1 license specifically for Snowflake and it failed completely to reverse engineer our database.  The parser threw errors where schema names were enclosed in double quotes (as needed) and when view columns had masking statements (as needed).  There were so many parser errors that I could not determine what else it could not do.  Sparx was nice enough but I recommend all interested to also email Sparx sales to increase the visibility of the fact that their product is useless for Snowflake and encourage them to come quickly up with a fix!