How to write a regular expression to match all patterns of these dynamic URLs?

[jagadeeshm]

I have front-end events with several dynamic uri patterns. I am trying to generate a report to summarize the average, perc95, perc80 for the response time by uri.

Some of the uri patterns include:

/market/us/asdfadf/home.do
/market/us/123123/home.do
/market/us/dfq3we/home.do
/market/us/dfq3we/home.do
/shop/us/buy.do
/shop/us/buy.do
/market/us/shop/dfq3we/home.do
/market/us/shop/dfq3we/home.do
/market/us/shop/dfq3we/home.do

I can certainly apply reg-ex pattern using sed and convert the alphanumeric characters to -- and display the counts as:

/market/us/--/home.do
/shop/us/buy.do
/market/us/shop/--/home.do

But my problem is I don't have list of all the patterns that match, so I can't write evals. I am wondering if there is a dynamic way of writing the regular expression so all patterns of these urls are identified.

Thanks!

[lguinn2]

I suggest that you look at the free app that can parse URLs/URIs for you: URL Toolbox
And on a related topic, here is the most recent app that can parse the user agent: TA-user-agent

Yes, you can write this yourself, but I prefer to use apps for this kind of functionality when possible!

[somesoni2]

You need to have some sort of rule/pattern for consolidating URIs, either for portion of URI you want to remove OR portion of URI you want to keep. Without that it doesn't seems feasible.