I am new to teradata.
Will the STATS be useful on below kind of data?
I think we can do sample statistics. Since sample statistics is better than default Random AMP statistics, which the optimizer will do if statistics is not available. Please correct me if I am wrong.
Stats would always be helpful to Optimizer as it gives the data Demographics i.e Unique values,nullabity,non nullabity in order to create the best plae time .
Sampled stats are recommended when the volume is too huge like 1 million, 2 million as full collect stats will be time consuming .
If no stats are collected then Optimzer will do a random amp sampling on the table in order to generate the plan for the query in which table is getting refered.
for more info. refer the below excellent link of carrie-
Depending on the release you are on and the data type of the column , sample stats may not be very useful.
If the column shown above is defined as character , it is larger than the 16 byte size that stats collection is most effective at.
Also if your table size is small , the sample is also not going to be the most effective.
I typicaly only used sample stats on USI's and UPI's (or anything fairly unique and not skewed) on very large tables (over several million rows) .