Have you heard of incremental refreshes? http://onlinehelp.tableau.com/current/pro/online/windows/en-us/help.htm#extracting_refresh.html
I may be misunderstanding what your issue is. Do you know why the refresh is failing? Are your refreshes spaced out? Do the refreshes all have the same priority? Are all the data sources hitting the same database? Are the data sources embedded in the workbook or are they published data sources on Tableau Server?
On a further note without looking at the workbook, 11 data sources seems an awful lot. Is there anyway to combine some of these into less datasources or to make multiple workbooks instead of the single one?
I know its a lot of questions but these are the things that people will want to know and might need to know to offer insight.
Trying to answer all your questions:-
The SQl query takes more than 10-12 hours as there are more than 18lakh records and the DWH server shuts down late night because of which the connection breaks and the query fails.
I need all 11 Data Sources as this workbook contains many dashboards, Yes all the data sources hit the same database, No these data sources cannot be embedded at the moment.
I tried to find the incremental refresh option but i did not see one, attached the screenshot of my tableau workbook, currently i am using tableau 9.2 desktop.
I wanted to know if there is something which can be done to save the time and effort?
When you select extract it gives you many options:
See below. This will only add new rows to the database. It will not look at
old rows and see if these have changed. Usually we do an incremental
extract once a day maybe twice or more depending and the full extract when
nobody is around like on weekends. If your queries are taking too long I
would suggest looking if you can minimise the workload in the query. That's
a long time even for a lot of records. Finally You can aggregate data for
visible dimensions as above so that you can minimise the number of rows in
your extracts. You would use this data source to answer questions and
build graphs that dont need the same level of data as you pull from the
I would be highly suspicious of 11 datasources that all would be needed but
then again I'm not an expert on your data
On 20 July 2016 at 11:50, Abhishek Rastogi <email@example.com>
image.png 20.9 KB