Allan_dk Tracker

When was data last updated

Allan_dk — Tue, 12 Apr 2016 13:08:38 GMT

Is it possible for a user that don't have access to SASVisualAnalyticsAdministrator to see when the data in LASR was updated.

A workaround is to add a variable last_updated when the data is created, but I think there must be another way around.

VA version 7.3

Sv: Category variable for coloring region maps

Allan_dk — Fri, 18 Dec 2015 14:17:09 GMT

Just moved to VA 7.3 and I am still not able to use category variables for coloring region maps, bubble map and coordinate maps.

Date based aggregation i SAS VA tables

Allan_dk — Mon, 30 Nov 2015 12:35:20 GMT

Is it possible to display aggregated date values in tables?

I have a dateset that contains data about users acces to VA. "Who do what when?"

[user, transaction date, event]

I would like to display information like

user, max_date, min_date

and give the user ability to filter data.

I can't calculate the max and min variables before I load data til LASR server.

That always semes to be the solution with SAS VA.

Re: SAS.IdentityGroups in Conditional Grants

Allan_dk — Tue, 07 Jul 2015 06:19:20 GMT

From thread https://communities.sas.com/thread/62416

You can activate the old editor on a table with a command like this

./sas-set-metadata-access -host servername -port 8561 -user USERNAME -password PASSWORD "Analytical LASR Data - Sandbox/SALES(Table)" -grant "SalesUsers":Read -condition '("rbs-"|| departmentname) IN ("SUB::SAS.IdentityGroups") OR "rbs-FullAccess" IN ("SUB::SAS.IdentityGroups")'

Works for VA 6.4

Re: Best practice for storing and sharing tables

Allan_dk — Fri, 27 Feb 2015 07:05:54 GMT

Thanks

The users find all there data through metadata/SAS Folders. I can teach them to go look at the serverlist and find data through libname, when they are looking for data created by other users, and not data from the dwh.

But it is inconsistant, to have data listed in two different areas. :smileyplain:

But I asked for best practice, and it is a better solution then what we are doing today.

So thanks for the help!

Re: Best practice for storing and sharing tables

Allan_dk — Thu, 26 Feb 2015 09:51:34 GMT

Made a libname where the team has read and write access.

Anthony is able to store data in the library.

But the data don't show up in SAS Folders.

Anthony has to run the Update Metadata on the library, before it shows up.

Is this still part of best practice?

For the user it seems like unnecessarystep.

And to delete data, the user has to run task "Delete data set and formats" and then Update Metadata again.

In almost every other programs you can just right click and select "Delete" or somthing that easy.

Re: How do you use SAS Theme Designer for Visual Analytics? SAS wants to know.

Allan_dk — Thu, 19 Feb 2015 11:55:19 GMT

I have used the theme designer to create a style that has a more clean look and uses the company colors.

Most important:

company logo at loading screen
logo and color of banner
white background
table header (olap table has to be changed directly in the xml-file)
colors on graphs data style

and then setting different elements to white (as background) to get a more clean look.

The theme designer (v.6.3) is easy to use except changes to the OLAP-table.

I would like to be able to

make changes to the OLAP-table directly in the designer
to change the log on screen
barchart.direction = vertical
remove barchart.axisline top/right
barchart.outlines

And it would be nice if the users have access to default company colors when he/she is making changes en the report designer.As for now the users have to type in color HEX codes in the Custom Colors pane.

The new theme:

va denmark.pdf - Google Drive

Best practice for storing and sharing tables

Allan_dk — Tue, 17 Feb 2015 13:53:41 GMT

What is the best way to let users share data in EG.

Scenario: Anthony and Benjamin, Carl, Daniel, Ethan

In EG Anthony uses som data on the server SASApp and creates at little query or two. By default this is stored as WORK.QUERY_FOR_CLASS_0001.

The next day Anthony, Benjamin, Carl, Daniel and Ethan wil use this data to do some more reporting, but Anthony is done for today.

Anthony then closes EG and it warns him "This project has references to temporary data......data will be deleted...".

Anthony has to sace the data because the source data changes over night and he can't recreate the dateset tomorrow.

Today we do the following:

Anthony uses "Task/Download datafiles to PC", and store i locally on the pc.

Anthony then sends it by email to Benjamin, C, D and E.

Benjamin uses "Task/Upload datafiles to PC". So do C, D and E.

As long as they work on the project they have to keep the files stored locally on the PC.

We use SAS 9.3 and EG 5.1.

Re: Range Slider Date Help

Allan_dk — Mon, 08 Dec 2014 13:16:51 GMT

There was a similar discussion earlier. But no good solution.

And remember, when you have multiple sections in a report, the user have to adjust the period on every section. The user will often miss this, and instead get the wrong information form the report.

Re: About Percentage

Allan_dk — Tue, 02 Dec 2014 10:00:49 GMT

I think you are asking for a way to calculate percent of column.

Right click on the column in your table.

Create and add/Percent of Subtotals...

Choose Percent of Column totals.

Re: Comparing one entity with the average of the population

Allan_dk — Tue, 04 Nov 2014 13:35:44 GMT

I have the same problem, and have not found a way around.

I don't think that Visual Analytics is intended for bench marking.

For now the best solution I could come up with is to display two separate graphs.

The left shows average of the hole population and the one to the right shows the department itself.

In the back there are two tables with the same data. But the table that I use on the right has row-level-security so that the user only get access til the departments own data.

Re: Automatic configuration of LASR Tables authorization

Allan_dk — Tue, 04 Nov 2014 12:16:47 GMT

Take a look at the batch-tools.

SAS(R) 9.4 Intelligence Platform: Security Administration Guide, Second Edition

The you are able to use commands like this

./sas-set-metadata-access -host servername -port 8561 -user USERNAME -password PASSWORD "Table" -grant "SalesUsers":Read -condition '("rbs-"|| departmentname) IN ("SUB::SAS.IdentityGroups")'

Re: Determine the current user in Visual Analytics

Allan_dk — Thu, 30 Oct 2014 09:10:43 GMT

In VA 6.1 you can write expressions like this

("rbs-"|| departmentname) IN ("SUB::SAS.IdentityGroups") OR "rbs-FullAccess" IN ("SUB::SAS.IdentityGroups")

If the user is member of the group "rbs-Sales" then they will have access to all rows where departmentname = 'Sales'.

I have a extra condition that allows users in the group 'rbs-FullAccess' to read all data.

There are other identity driven properties like SAS.Userid SAS.Personname, SAS.IdentityName that you can use.

BTW:

In VA 6.2 and 6.3 (and maybe even later versions) SAS have removed the option to write expression, and replaced the editor with a drag-and-drop editor, and you can not use identity driven properties.

There is a workaround.

You can activate the old editor on a table with a command like this

Re: DI-studio : SCD Type 2 - close out records

Allan_dk — Wed, 08 Oct 2014 17:48:21 GMT

Well, SAS Support, was able to help me.

"if you want only key values close out the first time encountered when using the option "Close out records not in source table", you would need to add the "User current indicator" under Change Tracking in addition to "User beginning and end dates". Adding the current indicator method will prevent closed keys from being retrieved during the subsequent executions of the SCD Type 2 process."

Re: DI-studio : SCD Type 2 - close out records

Allan_dk — Thu, 25 Sep 2014 06:33:44 GMT

Yes I know that i have all columns marked for detect changes. Including vtend and vtbegin.

In this case it makes sense - including vtbegin and vtend. These are just ordinary variables that I would like to keep track on.

Prompt on reportlevel

Allan_dk — Sun, 11 Oct 2015 02:30:42 GMT

When I have multiple sections in my report the user has to select the department on every section.

They sometimes forget this and spend time looking at section two that contains information from another department.

I can't always use linking between sections. Fx. when first section contains a filter on year and the second section shows data from all years.

Therefore I would like to have the opportunity to make af filter prompt on reportlevel.

Re: DI-studio : SCD Type 2 - close out records

Allan_dk — Mon, 22 Sep 2014 10:56:35 GMT

Here is the real data and some more dokumentation.

https://docs.google.com/document/d/1nuuzTusBcz1QgmVn_J_6A8Gc1vkeU8gdkj-lL9NqasI/edit?usp=sharing

Re: DI-studio : SCD Type 2 - close out records

Allan_dk — Thu, 18 Sep 2014 11:37:45 GMT

Beginning Date	vt_start	%sysfunc(datetime())
End Date	vt_end	'01Jan9999:00:00:00'DT

Except from that we use 9999 instead of 5999 you are correct.

Re: DI-studio : SCD Type 2 - close out records

Allan_dk — Thu, 18 Sep 2014 11:34:47 GMT

vt_end has 0:0:0 because it is just testdata.

There are no columns under SCD1

/* Close out any records that are in the xref table but not in the source. */

else if inXref and not inSort then

do;

ETLS_CLSDATE = DATETIME();

output work.etls_close;

end;

run;

DI-studio : SCD Type 2 - close out records

Allan_dk — Thu, 18 Sep 2014 09:05:21 GMT

In the transformations propeties/options it is possible to activate

Close out records not in source table: For records not in source table, assume they are to be closed out in target table.

When a record is missing in the source data the transformation set the validtime_end til %SYSFUNC(DATETIME()) and that is what I expect.

Before

ID	vt_start	vt_end	Value
1	01jan2014:0:0:0	15mar2014:0:0:0	A
1	15mar2014:0:0:0	31dec9999:0:0:0	B

After the record is found missing

ID	vt_start	vt_end	Value
1	01jan2014:0:0:0	15mar2014:0:0:0	A
1	15mar2014:0:0:0	17sep2014:0:0:0	B

But the next day the transformation runs it apparently updates the record in the target table.

ID	vt_start	vt_end	Value
1	01jan2014:0:0:0	15mar2014:0:0:0	A
1	15mar2014:0:0:0	18sep2014:0:0:0	B

I lose track of when the record was detected missing.

What to do?

DI-studio v4.6