sdoorneveld Tracker

Re: Check if Multiple Variables exist in a SAS Dataset

sdoorneveld — Fri, 28 Nov 2014 12:18:47 GMT

The method I always use is based on proc contents. It's almost similar to the method Reeza is describing, but less coding.

You run a proc contents on the test1 data-set. The keep variables you can use not only to check the varname but also if type, length or order matches.

proc contents data=test1 out=meta_test1 (keep=name type length varnum);

run;

Then you create a template file (or use the one that's registered in the metadata);

data meta_temp;

attrib Id length=8 format=8.

score1 length=8 format=8.

etc...

stop;

run;

Then the rest is the same as the proc sql; create table comparison from Reeza's code.

If you want to check of order, type, length, just add them to the join.

Re: Business rule based on data source

sdoorneveld — Mon, 26 May 2014 08:55:29 GMT

Finally found the answer to my problem. Thanks to this forum some people at SAS got triggered, and emailed me directly.

Emailed the answer to TS because they didn't know the solution and closed the track.

The solution is to create a DSN on the data management server that has the same name for every level, but addresses to the correct directory. This DSN name is used to create a data connection for the repositories which is used in the data source for business rules. This way, when you import the business rules on a level, it uses the correct folder via the DSN.

Re: Business rule based on data source

sdoorneveld — Wed, 14 May 2014 07:19:41 GMT

, I did send TS an email but I hoped to find some guidance here..

Business rule based on data source

sdoorneveld — Mon, 12 May 2014 09:11:38 GMT

When I'm defining a business rule (type: Compare a field to a filed in a data source), I'm experiencing the following problem:

In Data management Studio 2.4, I've created 3 repositories on one server so I have a DTA-roadmap. Each repository has it's own ODBC Data Connection. (DDMS, TDMS and ADMS)

I also created SAS Data Set connections, for each level, because I have business rules based of referential data stored in data sets. The main reason for this is that they're quite large and the content of the referential data is dynamic.

When I create a business rule, I have to choose a data source, say Main_lev1 (this is the SAS Data set connections for DDMS). The I choose the table and column.

When I export the business rule from DDMS and import it tot TDMS, it still searches for Main_lev1 and not for Main_lev2 (the SAS Data set connections for TDMS).

I can't use a macro variable when I define the business rule, I can only choose from a drop down box.

Has anyone encountered this problem? I'm running Dataflux combined with DI-studio, therefor the DTA-roadmap configuration.

Re: Oracle column length increase when metadata is registered to DI 4.2

sdoorneveld — Wed, 19 Mar 2014 09:36:39 GMT

When you define your libname, you can add serveral options. Then most important ones are DB_LENGTH_SEMANTICS_BYTE (defines if it's varchar2 of byte for character fields) and DB_CLIENT_MAX_BYTES (defines how much byte a single character must be).

here's an example:

LIBNAME ORA ORACLE INSERTBUFF=10000 READBUFF=50000 PATH=OSAS SCHEMA=SAS USER=SAS PASSWORD="blabla" DB_LENGTH_SEMANTICS_BYTE=NO DBCLIENT_MAX_BYTES=1;

DBCLIENT_MAX_BYTES=1 will prevent oracle for blowing up your column lengths by 4

Re: Decode EXEC.EX_DATE column in DataFlux repository

sdoorneveld — Wed, 12 Mar 2014 16:50:13 GMT

took me half a day to find out, but here is the result:

attrib new_date format=datetime20.;

new_date = intnx('dtyear',ex_date*3600*24,-60,'S');