Bash search logfiles

LogFile example

2012-09-16 19:06:33,450 ERROR [-59] JDBCException ORA-00001: unique constraint (X_UK1) violated
2012-09-16 19:31:39,319 ERROR [-60] JDBCException ORA-00060: deadlock detected (X_UK3) violated
2012-09-16 19:31:39,998 ERROR [-61] JDBCException ORA-00001: unique constraint (X_UK1) violated
2012-09-16 19:41:31,331 ERROR [-62] JDBCException ORA-00001: unique constraint (X_UK2) violated

First option is to grep the lines with ORA-00060:

$ grep ORA-00060 l.log 
2012-09-16 19:31:39,319 ERROR [-60] JDBCException ORA-00060: deadlock detected (X_UK3) violated

Find ORA error strip first part of the line.

$ grep ORA l.log | sed 's/.*ORA/ORA/g' 
ORA-00001: unique constraint (X_UK1) violated
ORA-00060: deadlock detected (X_UK3) violated
ORA-00001: unique constraint (X_UK1) violated
ORA-00001: unique constraint (X_UK2) violated

Search for ORA and count the time they occur.

$ grep ORA l.log | sed 's/.*ORA/ORA/g' | sort | uniq -c
   2 ORA-00001: unique constraint (X_UK1) violated
   1 ORA-00001: unique constraint (X_UK2) violated
   1 ORA-00060: deadlock detected (X_UK3) violated

Search with sed between two dates.

$ sed -n '/^2012-09-16 19:31/p; /^2012-09-16 19:32/q' l.log 
2012-09-16 19:31:39,319 ERROR [-60] JDBCException ORA-00060: deadlock detected (X_UK3) violated
2012-09-16 19:31:39,998 ERROR [-61] JDBCException ORA-00001: unique constraint (X_UK1) violated

String manipulation with sed

First example

$ echo "Hello world" | sed 's/o/i/g'
Helli wirld

Example with other seperator

$ echo "Hello world"  |sed "s@o@i@g"
Helli wirld

Example with slash forward

$ echo "Hello/ world"  |sed "s@/@.@g"
Hello. world

Example with slash backwards

$ echo "Hello\ world"  |sed "s@\@.@g"
sed: 1: "s@\@.@g": unterminated substitute in regular expression
$ echo "Hello\ world"  |sed "s@\\@.@g"
sed: 1: "s@\@.@g": unterminated substitute in regular expression
$ echo "Hello\ world"  |sed "s@m@.@g"
Hello\ world
$ echo "Hello\ world"  |sed "s@\\\@.@g"
Hello. world