KNOWLEDGE BASE

FLOAT Data-type Values Not Displaying as Expected When Connected to Snowflake


Published: 14 Aug 2018
Last Modified Date: 21 Aug 2018

Issue

When connecting to Snowflake from Tableau Desktop, values for fields designated as FLOAT data-types in the database do not match what is being displayed in data source.

Environment

  • Tableau Desktop
  • Snowflake

Resolution

Change the data-type in the Snowflake database to specify a precision and scale (i.e. NUMERIC(38,4)) - as long as the second number in the parentheses (representing 'scale') is less than 16 and greater than 0, the expected values should be returned in Tableau Desktop.  

Information on changing Data-types in Snowflake can be found here.

Cause

In Snowflake, FLOAT data types use double precision floating point numbers. This generally means that FLOAT data type fields do not have a precision/scale that cooperates with Tableau. Tableau may unexpectedly throw errors or improperly render any values whose scale is greater than 16 (For example, 7.9251349817081219.) as Tableau internally represents decimals with a scale limited to 16.

Did this article resolve the issue?