The configuration file used by data.get to determine how to access data from a given source. It is located in $DB/etc/$STATION/get.sources.$STATION.conf or in the default location $DB/etc/get.sources.conf. The default one is used if there is no definition for the given source in the station specific one.

For a given source, each program is called in the order it appears in the file. The first data it returns is assumed to mark the start of the time that it can generate data. That is, the programs should be ordered in such a way that the first ones access newer data.

Unless it is explicitly flagged as being able to handle multiple records concurrently, then each program is called once for each record data.get is attempting to access.

Each program is called with like “program $STATION $RECORDS $START $END $SOURCE”.


Lines beginning with '#' are treated as comments.



That data source being defined. Case insensitive.


That data source name may optionally be followed by a list of flags enclosed in parentheses and separated by semicolons.

The following flags are available:

  • localdata - Only call this handler for local (remote site data) access. This means that this program will only be called when data.get was called with the –localdata flag (generally only used by CPX2).
  • multirecord - Indicate that this handler can deal with multiple records at a time.
  • nofragment - Calls to this handler should not be fragmented based on cache or other record availability. This flag should be used on any handlers that can require locking beyond their explicitly called range (such as data.edit.get).
  • record:PATTERN - The record flag has a second part separated by a “:”. This handler will only be called for records that match the Perl regular expression (wrapped like /^$PATTERN$/, that is it must match the entire record name).


The program to call.