[BUG] Creating tables with all partition columns is disallowed, but dropping last non-partition column is allowed

[ebyhr]

Bug

Which Delta project/connector is this regarding?

• Spark
• Standalone
• Flink
• Kernel
• Other (fill in here)

Describe the problem

Steps to reproduce

Please include copy-pastable code snippets if possible.

1. create table default.test (c1 int, c2 int) using delta partitioned by (c2) TBLPROPERTIES('delta.columnMapping.mode'='name');
2. alter table default.test drop column c1;

Observed results

Creating tables with all partition columns (e.g. create table default.test2 (a int) using delta partitioned by (c2)) throws "Cannot use all columns for partition columns" error message, but dropping last non-partition columns succeeds.

Expected results

Throw exceptions when the target column of DROP COLUMN is the last non-partition column.

Further details

Environment information

• Delta Lake version: 2.3.0
• Spark version: 3.3.2
• Scala version: 2.12

Willingness to contribute

The Delta Lake Community encourages bug fix contributions. Would you or another member of your organization be willing to contribute a fix for this bug to the Delta Lake code base?

• No. I cannot contribute a bug fix at this time.