There are many 'convenience' filters that create very complex underlying patterns.
I have just added this line to the help:
Unlike a pattern match filter that matches everything and then throws it away, this filter is far more efficient, especially for large files, as it signals completion back to the input filter so only the first chunk of a multi-gigabyte file will ever get processed.
We are trying to move eventually to a position where Pro has just Mainframe functions as the extras.
The help file has been updated a lot to reflect this, but discrepancies will still crop up.
Please continue to let us know what you find - it has been very helpful to have your intelligent feedback. I also hope that the speed of updates suits you!