I have a table name called with the structure of "memp(a int, b varchar(20))"
In my flat file, i have 3 columns where i need to merge 2 and 3 columns together in the table memp by using multiload utility.
I have did in 2 ways, such as
1. ins memp(:c1,:c2||:c3);
Field c1 1 varchar(20),
Field c2 * varchar(20),
Field c3 * varchar(20),
Field c4 c2 || c3;
Here could you please let me know the internal process for these 2 method? and which one will be the best?
In the first case, three parameters are sent to Teradata and then the database does the concatenation. In the second case, the client on which MultiLoad is running does the concatenation and then sends two parameters to Teradata.
It depends on how you define "best". Often the first option will be faster, because Teradata AMPs will do the concatenation in parallel while the MultiLoad client will process serially. But slightly more data will be sent across the network and stored in the MultiLoad workfile.