Nfer event files are formatted as follows:
Each row is a pipe-separated list of either 2 or 4 values represents one event.
The first column is the event name. The second is the event timestamp (integer).
The third and fourth columns are used when the event carries data.
The third column is a semicolon separated list of data keys.
The fourth column is a semicolon separate list of data values in the same order
as the keys.
Example event file contents:
foo|123|x|2
bar|987|x;y|3;true
The result of reading such a file with this function is a dataframe where the
first column contains event names, the second contains event timestamps, and
subsequent columns contain data. The data is formatted with one column per name.
Events that don't carry some data key will have an NA value for that column.
Resulting dataframe of event:
| name | timestamp | x | y |
-------------------------------
| foo | 123 | 2 | NA |
| bar | 987 | 3 | TRUE |