EG usage advice - for adhoc data manipulation after filtering

[deleted_user]

Hi

We have a large migration project underway moving users from display manager across to Enterprise Guide. This has been largely successful, however there are a couple of sticking points for clients one of them is the lack of a 'Data/Where' option in EG to allow for users to subset a large sas dataset, and modify the data - in place - for ad-hoc data manipulation.

Whilst aware that EG will allow you to 'filter and query' a data set to subset it, this will produce another dataset in another location as output. This auxillary dataset can be modified in an adhoc manner... but not the original. Presumably the auxillary dataset that is created through this 'filter and query' task can be merged back in with the original, but this seems a retrograde step for users.

Are you able to advise if I am correct in that this is the only way to achieve equivalent ad-hoc data manipulation after subsetting in EG? If so, can you advise on the best way to merge the data back after it has been queried, and modified.

thanks in advance

[ChrisHemedinger]

Adam,

That's correct -- currently there is no user interface (like VIEWTABLE) to allow you to subset the data in place. All of the filtering steps in EG are recorded as steps in the project so that you can re-run them as a flow, and each step must result in another output table.

To save space, you can alter your query/filter step to create a VIEW instead of a TABLE. Also, if you want to really modify the data permanently in place, you can create a code step that performs a DATA step to do that work:

[pre]
data work.class;
set work.class(where=(age>14));
run;
[/pre]
We are looking at ways to make this better in a future relase so that you can apply adhoc filters without creating new tables.

Chris

[JohnH]

In EGuide, do a filter and query, subset your data, but click on options; options for this query; click on override preferences and choose data view. Then, you don't create the full dataset, you basically save the code and create rows of the data as needed. i like it because I can pretend that the data really exists, when it does not. It is a virtual dataset. You can read about data views in the online doc.