Hi,

See the attached for two options. One uses a table calc and requires Date to be in the view. The other is how I tend to do this sort of thing in Tableau, I use a subquery (in this case, I used Custom SQL) to generate another column for the latest date and then it's a simple Boolean evaluation.

Jonathan

Hi John,

Good question, here's my thinking:

- When using the table calc, Tableau is returning every date from the data source and computjng row-level and regular aggregate results for all the rows that are hidden by the table calc filter. This may not performant.

- Putting Date on the LoD increases the level of detail in the view. If you have other aggregates in the view, then they all need to take that into account, which can potentially mean a whole bunch of table calcs that have to be maintained.

Jonathan

Hi John,

I'm thinking that the direction you were going in is more complicated than necessary since it seems like all you're looking for is the % diff from prior for the latest date?

There are a number of ways to do this, I set up four options in the attached:

option 1 - using the Quick Table calc % diff from prior, and the TC filter for the latest date.

option 2 - same quick table calc, just using a LAST()==0 filter

option 3 - using the custom SQL data and a filter to just get the last two dates (so Tableau doesn't have to return all the dates from the data source), then is set up like option 1.

option 4 - using the custom SQL data and a row-level calc to do the % diff for the last two dates, does not require Date to be in the view.

Jonathan

It's true that most filters come first, but not filters based on table calculations: Here's a simplified order of operations:

(all three of these filters are processed in the data source)

Data Source Filters

Context Filters

Regular Filters

Computation of row level calculations

Computation of aggregate calculations

(the rest of these happen inside in Tableau)

Data densification

Data blending

Compute table calculations

Table calc filters

Filters on generated lat/long

Hiding

Within the three types of filters that happen in the data source, there's a sub-ordering:

Top & Conditional Filters

General & Wildcard Filters, and Range filters on dimensions

Filters on regular aggregates

So filters on (almost***) anything that is not a table calculation are applied within the query to the datasource, whereas filters on table calcs are applied after most all computation is complete. The almost*** is because there are a couple of situations with data blending where Tableau can do its own computation and and application of filters on aggregates (and maybe row-level, I'd have to go back to my notes to check).

The best post on this is http://community.tableau.com/message/139603#139603

Does this help?

Jonathan

You're welcome!

Hi

I am trying to download the file .. but cannot .. it is deleted or something ?

Thank you!

Hi

Can you please repost this file as an packaged work book I am unable to open it

Thanks

I have almost the same problem but I can't seem to find the solution. I need to identify which one is the latest Activity for each ID Activity.

This is an example of what I have:

This is what I need:

I know that if I create a calculated field with the max date for each ID Activity I can filter them and that's it, but how can I do that. I've tried several things but none useful to mention.

I hope you guys can help me! Barry PlumThe specified item was not found.Amy SchneiderThe specified item was not found.Erin Easter Elissa Fink

Thanks very much.

have you found a solution here ?

I have tried

If [date]={ fixed [id_activity]: max([date])} then true

else false

end

however didn't get to the req result.

Hi,

Use this,

{ FIXED [Activity]: Max([Date])}