The file size does not depend on the data source. If you want to reduce the file size then you need to apply joins .
Please go through this article :http://www.tableau.com/sites/default/files/whitepapers/designing-efficient-workbooks.pdf
I don't need join here, as I have only 1 data set here. But does performance depend on the data source or that will also be as same as this?
1 of 1 people found this helpful
I wouldn't say to reduce file size you need to apply joins - there are many ways to reduce the file size. Joining may actually increase your file size if it results in more columns coming into the dataset, i.e. a LEFT join.
The source of the data should not matter if all three sources return exactly the same dataset. If one source returns more columns than another, then it would result in a bigger file.
If you want to be really sure, go ahead and connect to each and create an extract and compare.
Performance will definitely depend on the source of the data. The extract speed will change depending on the source. Once it's extracted though, the source becomes a bit irrelevant and they will all perform the same.
Just to add a little more detail - you may see a slight file size difference between the three data source. Each data source reports its data types to Tableau in slightly different ways, and so you may not end up with exactly the same result. For instance, the legacy Excel connector always reported columns as double-precision and Tableau would never shrink that to an integer even if all values were integers. In your case, Teradata has relatively rich data types, text somewhat less and SAS the least. You may also see differences depending on what the data sources report as the collation of the data, especially case-sensitivity.
To know for sure, you'd have to try all three.
For tuning workbook size, the best thing to do is hide columns you don't need so they can be dropped from the extract.
For performance, see the article that was referenced earlier.