I work on Oracle DB , I would convert date at DB level only (Create a new View and then connect to Tableau)
Or make use of Custom SQL (instead of VIEW)
1 of 1 people found this helpful
Not know home much underlying data you have in the table but the best options would be to somehow convert the string into a valid date within SQL server.
You might be able to get away with creating a view on top of the table that could do that. The other one is to add a new column to the table and update it with a valid date.
This code will convert your you string to a valid date. Notice I needed to add a ":" into the -0800 to make it -08:00. This is so the conversion function would work to otherwise it is an invalid conversion.
declare @Originalstr varchar(40) = '2018-02-26T14:38:43.997-0800';
declare @str varchar(40) = left(@originalStr,len(@Originalstr) -2) + ':' + right(@Originalstr,2);
select convert(datetimeoffset(7),@str,4) ValidDateTime
Now once the date is valid you should be able to use this within Tableau to get your extract smaller.
Hope this helps.
Thank you Jeff!
I have an additional question...
Tableau seem's to be using UTC as the value to check against on my incremental refresh. This doesn't make sense since my server is running on UTC -8 and all timestamps from the system are also UTC -8. So I'm not getting any "new data" since it's looking for a time that is in the future!
Ignore that last post, I was able to figure out my problem and it had nothing to do with the Tableau Server Extract.. purely an oversight on my part.