2 thoughts on “1.1.2. Choose between live connection or extract”
Saileela
Hi.. I think this statement is not correct as we are able to count distinct in Live connection also.
Extracts offer additional Tableau functionalities, such as computing Count Distinct, which may not be available with live connections. and I cannot agree on the below as I have seen many reports go really slow if the data reaches certain million rows itself.
Extracts can handle massive amounts of data, even reaching billions of rows
The Count Distinct function is/was unavailable in some drivers used in live connections (ex: Microsoft JET Database engine) but as far as I know many of these drivers have been replaced with newer ones that can use the Count Distinct. This doesn’t make the statement incorrect since there are still functions not available in live connections (count distinct was used as an example) like MAKETIME.
Regarding the amount of data statement – yes reports can become slow with a massive amount of data and a developer should always need to optimize as much as possible the model. But extract is the option to choose vs live in this case since a live connection would most likely not be able to handle the amount of data, especially if there are complex calculations involved.
Hi.. I think this statement is not correct as we are able to count distinct in Live connection also.
Extracts offer additional Tableau functionalities, such as computing Count Distinct, which may not be available with live connections. and I cannot agree on the below as I have seen many reports go really slow if the data reaches certain million rows itself.
Extracts can handle massive amounts of data, even reaching billions of rows
Hi,
The Count Distinct function is/was unavailable in some drivers used in live connections (ex: Microsoft JET Database engine) but as far as I know many of these drivers have been replaced with newer ones that can use the Count Distinct. This doesn’t make the statement incorrect since there are still functions not available in live connections (count distinct was used as an example) like MAKETIME.
Regarding the amount of data statement – yes reports can become slow with a massive amount of data and a developer should always need to optimize as much as possible the model. But extract is the option to choose vs live in this case since a live connection would most likely not be able to handle the amount of data, especially if there are complex calculations involved.