Skip to main content

Current advice on partitioning on a non-date field calls for creating a partitioning “bucket” using field transformations. While attempting to follow the instructions outlined here I encountered a problem. 

Partitioning works by creating a calculated field in the valid table that has the capacity to handle NULL values if the partition template is so configured. However, that same field, along with the calculated default value, is also present in the Raw table. 

This means that when anyone following TimeXtender’s guidance for non-date transformations executes the table, that execution will fail with the error “Cannot insert the value NULL into column ‘DW_PartitionKey’”

Is this a bug, or does the documentation for creating partitions need to be updated? 

Ah! A follow up to this: 

This problem ONLY occurs if you have custom data in the table. I believe this is because the custom data algorithm attempts to insert a NULL value into the caculated field. 


Hi @ekw9 

Can you please send some screenshots of the table and the partition settings used?


Hi @ekw9 just following up on this - could you please send some screenshots of the table and the partition settings used?


Reply