The Teradata Database channel includes discussions around advanced Teradata features such as high-performance parallel database technology, the optimizer, mixed workload management solutions, and other related technologies.
I was going through columnar concept of TD 14.10. Columnar is still inactive on my system. So thought of putting these questions on forum
Hope these are relevant J
Will fallback, Secondary Index Sub tables of columnar table be columnar in nature?
Will these tables inherit properties like NO AUTO COMPRESS?
If NO AUTO COMPRESS is ON for all the columns in columnar table can we use BLC on fallback tables?
Can we use MVC on columnar table columns?
Is it possible to use MVC on columns in hybrid ROW and COLUMN storage? ( Column release_date, sold, producer, lyric)
CREATE TABLE album_artist_hybrid_row_col_noac (
artist CHARACTER(35) NOT NULL NO AUTOCOMPRESS,
album CHARACTER(50) NOT NULL NO AUTOCOMPRESS,
ROW (release_date DATE ,
sold INTEGER NOT NULL,
producer CHARACTER (40) ,
lyric LONG VARCHAR))
NO PRIMARY INDEX
PARTITION BY COLUMN;
What will be best practicies for collecting stats on Colamanar tables?
Is it advisible to use NUSI on column partitin tables? As NUSI will be ALL amp operation and as columnar tables are NoPI tables it will also be all AMP operation.
Are there any new joining strategies for joining between ROW partition table and Column partition tables? Or column partition tables will always we converted to row base table while getting into spool? And if so how stats on columnar tables will be helpful in this scenario?
If there is row base table and it is having data spikes ( uneven data distribution), then if we are creating NoPI table using CREAT TABLE AS, NoPI table will have same data spikes. But if we are creating Columanr table will it be same, will data spikes be there in Columnar table? (In some blogs I read that column containers will have same number of rows that’s why this question)