When creating or using view descriptors, follow these guidelines to minimize the
use of
CA-Datacom/DB and your operating system resources
and to reduce the time
CA-Datacom/DB
takes to access data.
-
Select only the fields your program
needs. Selecting unnecessary fields adds extra processing time.
-
Specify the order in which records
are presented to SAS (with a SORT clause or a SAS BY statement) only
if SAS needs the data in a particular order for subsequent processing.
The SAS BY statement
issues an ordering clause to CA-Datacom/DB so that CA-Datacom/DB does the sorting using its system resources. This SORT clause overrides any existing
SORT clause for the view descriptor. If you decide to use a SORT clause or a SAS
BY statement, order by a key, which is indexed, when possible. (For help with determining
which fields in a table
are indexed, see your DBA or the table's creator.)
As an alternative to
using a SORT clause, which consumes CPU time each time you access
the CA-Datacom/DB table, you could use the SORT procedure with the OUT= option to create a sorted SAS
data file. This is a better approach for data that you want to use many times.
-
If a view descriptor describes a large
CA-Datacom/DB table and you use the view descriptor often, it might be more efficient to extract
the data and place them in
a SAS data file. (Even though the extracted data file is very large, you need to
create it only
once. Also, the extracted data does not reflect any subsequent updates to the table.)
See
Performance Considerations for more information about when
it is best to extract data.
-
Specify selection criteria to retrieve
a subset of the records CA-Datacom/DB
software returns to SAS, where possible.
-
If you use a Default Key, the interface view engine uses an
index read instead of a sort if it can. Index reads are faster, but not always possible.
For example, an index read is not possible if you specify multiple sort keys, multiple
WHERE clause conditions, or a WHERE clause condition with a column that is not a key.
-
Omit the KEY statement if you do not need a certain order and you want to retrieve
the data sequentially. Otherwise, you might cause an unnecessary sort. PROC FSBROWSE,
FSEDIT, and FSVIEW automatically use random access and require a value in the Default
Key field.
-
You can provide your own
URT that is fine-tuned for your applications.