How to tell the difference in images?

[torturedata]

I have a new problem when analyse the log file.

Some images are embedded in a page, when the page is visited, all images included in that page are automatically requested by that page, and recorded in the log file.

Some images are not in the page, there is a link provided in the page, when the link is clicked, the image is requested and show up in a separate page.

Considering that the images in the second case maybe more important than the first kind, can we discriminate this two cases only from the log file?

[cfx211]

The problem is that log files record a row for every request that you are making and this is done at the hit level, not that the page level.

For instance if the same image is served as a part of the page, and on its own from a link off of a page, they are both going to get a row in your log file.

There are a couple of ways around this, either give the same image two names and serve one as a part of the page and the other as a stand alone, or change the link so that there is something useless added on the end like?image=full.

If you add something to the end of your link, you can then sort your traffic by referring URL. The one that is a part of the page will have the page name as the referrer, and the one coming from the link will have the page name with the?image=full on it.

[torturedata]

Thanks cfx211!
I think they are useful for a live server.
The data I'm using is not from a live server. So it seems there is no way to discriminate those two cases from the old log file.