Help, possible script?

[utterdisgust]

Hey, I have a question and I really don't know how to approach it...

I have text files which contain logs of certain events. Within these files I would like to group like events (such as ADD, UPDATE, DELETE or whatever it might be) and output these to respective text files. Every event has a time stamp which I would like to keep and output into my text files. Does anyone know what I could do to achieve this? I am new to running commands via the terminal and don't have the most extensive knowledge of commands to run. I do know however understand that I will need to use piping to achieve what I want. I am running CentOS and using BASH. Can anyone help?

[burschik]

You can do this with sed, awk or perl, for example. In sed, you could use something like this:

sed '/pattern1/wFile1
     /pattern2/wFile2
    ....' inputfile

[utterdisgust]

If I was to do this however, would I have to predetermine the patterns? Here is maybe a clearer representation of I want to achieve..

I want my command (or script, whatever) to...

1. Open the log file
2. Search document for potential patterns
3. If no like patterns exist, exit and leave log file as is
4. If like patterns exist, output grouped patterns to pattern named files in a new directory.

Like patterns exist if there are matching word(s). For example in the case of my previous example, if it is seen that there are two lines that contain UPDATE then they contain the like pattern of UPDATE.

[pavlo_7]

Let's say file.log contains:

ADD, UPDATE, DELETE or whatever
whatever ADD, UPDATE or DELETE
DELETE
UPDATE 
or
whatever
whatever
whatever
pattarn
patternnnnn.nn,--
more stuff
pattern
apattern
bbpattern
bbbpattern
another bbbpattern
pattern

The script might be something like this:

#!/bin/bash

echo -n "what to search:"
read pattern

if [ -z "`grep $pattern file.log`" ]; then
 
   echo Nothing found. Exiting...;
exit 0
     
     else
        if [ ! -d found-patterns ]; then
                 mkdir found-patterns;
        fi 
                grep "$pattern" file.log| sort -u >> .tmp-file;
                        for i in `cat .tmp-file`; do
                                if [ ! -d found-patterns/$i ]; then
                                        mkdir  found-patterns/$i;
                                fi
                                        echo "`grep "$i" file.log`" >  found-patterns/$i/$i.txt
        
                        done
				rm -f .tmp-file
				echo "Check found-patterns directory"
fi

Hope that helps.