I wrote a script in Python to do the work for me. It's called remove-orphaned-raw-images.py and I published it on Github.

Basically it iterates over all the files in a given folder and moves orphaned raw images (in my case *.CR2 files with no matching JPEG) to a backup folder. Optionally you can tell the script to actually delete the files.

Here is an outline of the algorithm:

• Get a list of all the files in the selected directory.
• Sort those files into RAW and JPEG files (append them to separate lists).
• Check for each item in the RAW images list that a match exists in the JPEG list;
  if not, append this image to a list of orphaned raw images.
• Move the images in the list of orphaned images to a backup folder
  (or directly delete them if explicitly wanted).

The tool will tell you how to use it when run with the help option -h on the command line.

This problem also occured to me, which is why I wrote this tool. I'm using my DSLR to take JPEG or RAW+JPEG images, never only RAW. When sorting out blurry or otherwise bad shots, I use the JPEGs to quickly take a look at them and delete the bad ones. This leaves me with left over RAW images of which I deleted the matching JPEGs (for a reason).