Hi,
I have a lot of Oracle DB package bodies in files and need to replace the names of internal called procedures/functions.
A small example:
suppose we want to change InsertCustomer to InsCust in a procedure like this:

PROCEDURE InsertCustomers () AS
BEGIN
SELECT InsertCustomer() FROM DUAL; -- this should be replaced
...
SELECT someOtherPackage.InsertCustomer() FROM DUAL -- this not
END InsertCustomers;

I need a restriction for sed, which excludes the lines 1,5 and 6 but in case insensitive mode (END can also be end or End, ...). There's an option /i in sed for the case insensitive mode, but that only applies for replacement and not for the restriction pattern.

I use following command, but that excludes only lines with PROCEDURE, FUNCTION, END in upper case and 'anything.InsertCustomer':

sed '/^\(PROCEDURE\|FUNCTION\|END\|.*\+\.InsertCustomer\)/ !s/InsertCustomer/InsCust/i'

I also need to have the ability to exclude lines like this:
anything.insertcustomer
anything.INSERTCUSTOMER
End
procedure
...

Thank you for any help.
Filip

According to this, gnu sed has an 'I' flag for case-insensitivity that can be used after regex expressions. I've tried a couple of simple tests and it seems to work both in the matching and in the replacement expressions.

Hi David,
thank you for your answer. You are absolutely right about the I flag, but as you said it works for matching and replacement, but not for the restriction regex.

Take a sed command like this:

sed '/restriction_regex/ s/match_regex/replace_regex/I'

The I flag will work on match_regex and replace_regex, but the restriction_regex will not be executed in case insensitive mode.

It's certainly working for me.