Identifying data sources and targets
On z/OS, data sources and targets that exist as datasets are identified by DDNAME. Sources and targets must be defined by DD statements in the execution JCL before a map can be executed.
When you create a map to be run using the Command Server for z/OS Batch, define source and target names associated with input and output cards. If a map is tested on a Windows platform before being executed on a z/OS platform, data source and target names typically use the Windows file naming convention of drive:path\filename.extension. When a map executes on z/OS Batch, source and target names are converted to valid z/OS DDNAMES by removing drive, path, and extension, converting _ characters to # characters, and translating the remaining characters to uppercase.
Consider the following example. A map is to be executed on z/OS Batch. It has one input and one output-both of them sequential datasets. The map was developed on a Windows platform; for testing purposes the data source was initially defined as C:\My\Data\In_file.txt and the data target was defined as F:\Outfile.dat. When the map executes on z/OS Batch, source and target names are converted to DDNAMES. The data source name becomes IN#FILE and the data target name becomes OUTFILE. The execution JCL must contain the following DD statements:
//IN#FILE DD DSN=MY.INPUT,...
//OUTFILE DD DSN=MY.OUTFILE,...
The DDNAME associated with a data source or target may be changed at run time using a card override on the command line passed in the PARM statement. Input and output card override commands are discussed in z/OS Batch Execution Commands .
The use of instream datasets for data sources is supported. If an instream dataset is used, ensure that no sequence numbers exist in columns 73 through 80; sequence numbers will be interpreted as data by the map. The GETANDSET function may not be used with instream datasets.