Creating and Loading SPD Engine Files |
You can convert existing default Base SAS engine data sets to SPD Engine data sets using the following methods:
PROC COPY
PROC APPEND
Base SAS File Characteristic | Conversion Result |
---|---|
Indexes | Rebuilt in SPD Engine (in parallel if ASYNCINDEX=YES) |
Default Base SAS engine COMPRESS=YES |CHAR |BINARY * | Converts with compression if the data file is not encrypted |
Default Base SAS engine ENCRYPT=YES * | Converts with encryption |
Integrity constraints | Dropped without ERROR |
Audit file | Dropped without ERROR |
Generations file | Dropped without ERROR |
* If the default Base SAS engine file has both compression and encryption, the compression is dropped, but the encryption is retained. SAS retains the security of the data set instead of the compression.
Converting Default Base SAS Engine Data Sets Using PROC COPY |
To create an SPD Engine data set from an existing default Base SAS engine data set you can simply use the COPY procedure as shown in the following example. The PROC COPY statement copies the default Base SAS engine-formatted data set LOCAL.RACQUETS to a new SPD Engine-formatted data set SPORT.RACQUETS.
libname sport spde 'conversion_area'; proc copy in=local out=sport; select racquets; run;
Even though the indexes on the default Base SAS engine data set are automatically regenerated as the SPD Engine indexes (both .hdx and .idx files), they are not created in parallel because the data set option ASYNCINDEX=NO is the default.
If an SPD Engine data set is encrypted, only the data component files are encrypted. The metadata and both index files are not encrypted.
Converting Default Base SAS Engine Data Sets Using PROC APPEND |
Use the APPEND procedure when you need to specify data set options for a new SPD Engine data set.
The following example creates an SPD Engine data set from a default Base SAS engine data set using PROC APPEND. The ASYNCINDEX=YES data set option specifies to build the indexes in parallel. The PARTSIZE= option specifies to create partitions of 100 megabytes.
libname spdelib spde 'old_data'; libname somelib 'old_data'; proc append base=spdelib.cars (asyncindex=yes partsize=100) data=somelib.cars; run;
Copyright © 2010 by SAS Institute Inc., Cary, NC, USA. All rights reserved.