Shell Programming and Scripting

Sorry if this has been posted before, I searched but not sure what I really want to do.

I have a file with records that show who has logged into my application:
2003-03-14[09:26:51]:I:root: Log_mesg: registered servername:userid. (more after this)

I want to pull out the userid, date and time into seperate fields to format into a file for a report. I figured out the -F: will let me seperate out the fields by : but I can't get a format for the brackets. Do I need to use a substring?
Is there a way to tell awk to see mutliple seperators or do I have to break this down into statements?

To go further into what I am doing. I am creating application specific auditing report (this level of reporting does not exist in the application).. logging all users that log onto our server, then by using the license log, identifying all users that logged onto the application and what time. I further have to accumulate this into reporting with monthly totals...etc etc etc.

I do not have control over accounting features...

Digital unix 4.0D (yes, I know its very old....)


Any help would be appreciated.

Thanks!

Normally, I'd post a ksh solution. But this morning's other questions have put me into a sed mood.

Suppose that:
2003-03-14[09:26:51]:I:root: Log_mesg:...
was to become:
2003-03-14:09-26-51:I:root: Log_mesg:...
This would make the problem much easier in any language. So all we need to do is use a sed script that will change something like:
[09:26:51]
into
:09-26-51
which is something that sed can do with a little effort...

Code:

sed 's/\[\([0-9][0-9]\):\([0-9][0-9]\):\([0-9][0-9]\)\]/:\1-\2-\3/'

Great! thinking out of the box!!!