Why does WAVE tell me `Record ... is unavailable'?

In common with other WFDB applications, WAVE searches for its input signals and annotations in directories named by the WFDB environment variable (see the discussion of the WFDB path in the WFDB Programmer's Guide). If WAVE is running remotely, remember to set WFDB on the WAVE host (a default can usually be set for C-shell users by `source /usr/bin/cshsetwfdb', or for Bourne shell, Korn shell, and `bash' users by `. setwfdb'; don't omit the `.'). The directories named by WFDB are those on the WAVE host; thus (unless your WFDB files are already on the WAVE host, or are accessible to the WAVE host via HTTP or FTP - see the next question) you must either transfer them to the remote machine, or NFS- or RFS-mount the directory in which they reside onto a suitable point in the file system of the WAVE host. Furthermore, remember that output annotation, log, and print files will be written to your current directory on the WAVE host.

To get the pathname of a WFDB file, use `wfdbwhich' (see wfdbwhich(1) for details). For example, the output of `wfdbwhich header 100s' is the pathname of the header file for record 100s (usually `/usr/database/100s.hea'). If `wfdbwhich' can't find the file, neither can WAVE (or any other WFDB application). In this case, you will need to check and correct the value of the WFDB environment variable. In unusual cases, the WFDB path may be correct, but you may not have permission to read the file or one of the directories in the path to the file; in such cases, you will need to get the file's owner or your system administrator to give you appropriate permissions.

George B. Moody (george@mit.edu)