-
Notifications
You must be signed in to change notification settings - Fork 48
SDMX Connector for SAS
The SAS Connector for SDMX.
All the instructions that follow refer to linux, but they can be easily adapted to Windows installations.
The only part of the connector that needs to be built is the Java library. Instructions can be found in the java section
NOTE: building the java library is not strictly necessary. A jar file, ready to be used, can be retrieved from the R package
-
Clone the repository or download a the full project in zip format and unpack it. Get the SAS files from the project SAS directory.
-
The SDMX functions are standard SAS macros. They can be loaded manually every time you want to use them, or you can set up an autocall library. Some more info can be found in the SAS Documentation
-
Build the java library or get it from the R package inst dir
-
The SDMX Java library needs to be added to the CLASSPATH that is configured in SAS. This can be either achieved by modifying the system CLASSPATH or by changing the SAS configuration. Some more info can be found in the SAS Documentation
-
If necessary, override the default configuration by means of a custom configuration file. This is particularly necessary if you want to run the connector in an environment that connects to the internet by means of a proxy.
Calling the graphical SDMX helper:
%sdmxhelp;
Get data (and, optionally, metadata):
single time series, freq=A, CURRENCY=USD
%gettimeseries(provider="ECB", tsKey="EXR.A.USD.EUR.SP00.A");
multiple time series, FREQUENCY=A (annual), all currencies
%gettimeseries(provider="ECB", tsKey="EXR.A.*.EUR.SP00.A");
get data and metadata
%gettimeseries(provider="ECB", tsKey="EXR.A.*.EUR.SP00.A", metadata=1);
NOTE: data results will be stored to a sdmxdata dataset. If enabled, metadata results will be stored to sdmxmetadata