I'm interested in using the columnar feature, and after doing some intinial research, I found that Teradata automatically determines if a column should be row or column format. I was hoping to get more information on how Teradata actually determines this, and if there is a certain way to enable this, or if the DDLs have to be in a certain format?
I did find many resources online with sample DDLs. In some, after the PARTITION BY COLUMN statement, the columns are specfied, and then ROWs are specfied. In others, they are not specified at all:
I would suggest reading the manuals, there should be lots of details in the DDL and the Database Design manuals plus the Columnar Orange Book.
Teradata does not automatically determine if a table should be row format or column format. That needs to be specified by the DBA at physical table definition time. The examples you provide show a couple of the options, the second is the most simple version making every column column storage. The first is a more complex example which shows a mix of column store and row store in the same table.
What Teradata does do automatically is to determine what compression algorithms to apply to the data in column stored columns (unless you tell it not to with No Auto Compression).