Best Of
Re: Averaging time/returning only whole numbers of a fraction number
@Jbrorby If you wrap a FLOOR() function around your entire formula, it will return the closest whole number without going over
Re: Transaction Deadlock workbench error
I haven't been using the ODBC driver myself. “Transaction was deadlocked” happen when 2 or more sessions are trying to lock a resource which has already locked by another session. The sessions cannot complete and SQL Server has to intervene. It gets rid of the deadlock by choosing one of the sessions as a victim and kills it allowing the other session to continue. The killed session is rolled back.
Re: Using partitioning in ETL or Workbench
id be very carefull with write backs we found it was not copying rows where the keys were duplicates.
column1 and column2 are the key identifier columns, column 3 new values to bring
ie old dataset
column 1 | column2 | column3 |
|---|---|---|
Date1 | AAA | 23 |
Date2 | BBB | 23 |
Date1 | AAAC | 24 |
new dataset
column 1 | column2 | column3 |
|---|---|---|
Date1 | AAA | 27 |
Date2 | BBB | 23 |
Date1 | AAA | 24 |
Date2 | BBB | 23 |
from the new dataset 1st column whould not copy over because partitiuon thought it already existed.
Maybe more documentation is required
Re: Using partitioning in ETL or Workbench
Partitions are used to replace a group of records typically. If you have some records being replaced for a date then pull in all records for that date and use the date as your partition key. This will replace all records with that date.
Alternatively if records aren’t being deleted you can utilize UPSERT in workbench to update just that single record based on the id
Re: Is there a reason why you cannot change the legend labels in a Period over Period chart type?
@jakebutterfield This may be able to work with variables and for that, you need to create two variables "Current year" and "Previous year" and then put them along with the BM to create a line graph.
(CASE WHEN YEAR(Date) = CY*1 THEN IFNULL(Values)`,0) ELSE 0 END)
(CASE when YEAR(Date) = PY then (IFNULL(Values)` ,0) ) else 0 end)
Re: Is there a reason why you cannot change the legend labels in a Period over Period chart type?
Looks like the series that you were using is Year. I would try creating a calculated field. Assuming you don't have any future dates in your data you could do something like this:
CASE
WHEN
DATE_SUB(DATE_SUB(CURDATE(), INTERVAL 12 MONTH), INTERVAL DAYOFMONTH(CURDATE())-1)
>DateField
THEN 'Current'
WHEN
DATE_SUB(DATE_SUB(CURDATE(), INTERVAL 24 MONTH), INTERVAL DAYOFMONTH(CURDATE())-1)
>DateField
THEN 'Previous'
END
You could then place this in your filter and select 'Current' and 'Previous'. Then use this field as your series as well. You can change the "Current" or "Previous" in the formula to whatever you want the label to say
Re: Is anyone having issues with filters using selectors
@alliu @JuanMiguel I just ran into the same thing but was able to resolve it by adding an optional group by field in the filter card
Re: How to build filters with images or icons
@GrantStowell - Thanks for reposting the videos
Re: Updating a User
Hey Ashleigh, I've seen this a few times if the new email matches a deleted user's email. We've also seen it happen lately if the user doesn't have some other attributes set (or if they're invalid). The browser Network tab might be able to give you more details on the error, or you can send the TOE from the response to support and we can investigate further.
