On Jan 29, 9:19 pm, j...@STANFORDALUMNI.ORG (Jack Hamilton) wrote: > On Jan 29, 2010, at 8:48 am, Fehd, Ronald J. (CDC/CCHIS/NCPHI) wrote: > > > sashelp.views are deprecated > > Where does SAS Institute say that sashelp views are deprecated? >

I never heard that term per se, but it's clear from the V9 documentation that they encourage use of SQL (using Tables) over non- SQL (using Views). Back in the V6 days, when the Tables were introduced, Nancy Michal (now Nancy Cole) and I wrote a paper describing the structure and uses of some of the Tables. We also did some comparisons of efficiency running VMS, Unix, Windows, and OS/2, comparing timings of creating a report from CATALOGS using SQL, a DATA step, PROC PRINT, etc. The SQL-based scenario (create a WORK dataset, then use PRINT) ran, on average, in 1/4 the time of the next-fastest approach.

Here's an excerpt from the V9.1 documentation:

DICTIONARY Tables and Performance

When you query a DICTIONARY table, SAS gathers information that is pertinent to that table. Depending on the DICTIONARY table that is being queried, this process can include searching libraries, opening tables, and executing views. Unlike other SAS procedures and the DATA step, PROC SQL can improve this process by optimizing the query before the select process is launched. Therefore, although it is possible to access DICTIONARY table information with SAS procedures or the DATA step by using the SASHELP views, it is often more efficient to use PROC SQL instead.

For example, the following programs both produce the same result, but the PROC SQL step runs much faster because the WHERE clause is processed prior to opening the tables that are referenced by the SASHELP.VCOLUMN view:

data mytable; set sashelp.vcolumn; where libname='WORK' and memname='SALES'; run;

proc sql; create table mytable as select * from sashelp.vcolumn where libname='WORK' and memname='SALES'; quit;

Note: SAS does not maintain DICTIONARY table information between queries. Each query of a DICTIONARY table launches a new discovery process.

If you are querying the same DICTIONARY table several times in a row, you can get even faster performance by creating a temporary SAS data set (with the DATA step SET statement or PROC SQL CREATE TABLE AS statement) with the information that you desire and run your query against that data set.