There must be a reason. Can you provide results of the following queries against both sources?
select count(distinct JOIN_FIELD) dcnt, sum(cnt) cnt, AVG(cnt) avg from (select JOIN_FIELD, count(*) cnt from the_table group by JOIN_FIELD) x
How many rows do you have in Teradata table?
In order to perform the join, Tableau has to suck in both tables and progress dialog you pasted could be from that part of the flow.
I got millions of rows in Teradata but I am expecting a result of just 16
rows. Result should be the same with the direct query result in Teradata...
On Sep 13, 2016 5:48 AM, "Dmitry Chirkov" <email@example.com>
16 rows is a result of intersection I assume, correct?
In order to filter the data, Tableau has to bring in the data first. This is just how this feature works.
If you are not providing any additional filters then Tableau has no other choice but to bring all those millions of rows locally for further processing.
if possible, can you join your csv file with a custom query over Teradata limitting the number of rows retrieved?
Ivan, is this a question for me or suggestion for Christopher?
it was mor a suggestion for Christopher
Below is the update from Support team.
I got the update from our development team about the "issue" of extracting
unexpected rows of data when joining CSV and Teradata in Tableau Desktop 10.
This is an limitation of the current version: all cross-data sources joins
are local to the TDE. We have to pull all data locally first. If you let
it finish, you will probably end up with the 16 rows. In future releases,
we are looking to make improvements, and perhaps that can be done remotely.
Chris C. Angeles
Enterprise Data Office, Globe Telecom
02-5871987 | 0917-6880102 | firstname.lastname@example.org
On Mon, Sep 26, 2016 at 2:31 PM, Ivan Monnier <email@example.com>