One regex works but not the other? (ksh)

I'm new to sed and regular expressions, and I'm hoping
someone can
explain to me why one regular expression works in ksh and
not
another...

I have a bunch of files containing CREATE TABLE sql
statements like so:

CREATE TABLE db.table {
         KEY_COLUMN           VARCHAR(20)
        ,KEY_COLUMN_02        CHAR(20)
        ,KEY_COLUMN_03        TIMESTAMP(6)
        ,NOT_A_KEY_COLUMN     TIMESTAMP(6)
        ,CREATE_MODULE        VARCHAR(1024)
        ..........
};

The format of this file is very well-defined. I want to
match any line
with both:

1. Column name of 'KEY_COLUMN' (not 'xyz_KEY_COLUMN' or
'KEY_COLUMN_abc')
2. Type CHAR or VARCHAR (not TIMESTAMP or DECIMAL(9,2))

My first attempt at a regular expression which matches these
lines was:
           
'<space>*,?KEY_COLUMN<space>*(VAR)?CHAR([0-9]*)'

The above expression does not work. What does work is:
           
'<space>*[<space>,]KEY_COLUMN<space>*[A-Z]
*CHAR[0-9]*)'

The first one makes a lot more sense to me:
            '<space>*[<space>,]'
seems like a confusing way of saying 'any number of spaces
followed by
an optional comma', which is what I thought
'<space>*,?' meant

and

            '[A-Z]*CHAR'
does what I need it to do, but I still don't understand
what's wrong
with '(VAR)?CHAR' As I said, I have an expression that
works, but I
don't feel like I learned anything - I want to figure out
why what I
had didn't work. Both failures did use question marks. Any
advice?


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the
Google Groups "Regex" group.
To post to this group, send email to regexgooglegroups.com
To unsubscribe from this group, send email to
regex-unsubscribegooglegroups.com
For more options, visit this group at http://grou
ps-beta.google.com/group/regex
-~----------~----~----~----~------~----~------~--~---

anon wrote:
> I'm new to sed and regular expressions, and I'm hoping
someone can
> explain to me why one regular expression works in ksh
and not
> another...
>
> I have a bunch of files containing CREATE TABLE sql
statements like so:
>
> CREATE TABLE db.table {
>          KEY_COLUMN           VARCHAR(20)
>         ,KEY_COLUMN_02        CHAR(20)
>         ,KEY_COLUMN_03        TIMESTAMP(6)
>         ,NOT_A_KEY_COLUMN     TIMESTAMP(6)
>         ,CREATE_MODULE        VARCHAR(1024)
>         ..........
> };
>
> The format of this file is very well-defined. I want to
match any line
> with both:
>
> 1. Column name of 'KEY_COLUMN' (not 'xyz_KEY_COLUMN' or
> 'KEY_COLUMN_abc')
> 2. Type CHAR or VARCHAR (not TIMESTAMP or DECIMAL(9,2))
>
> My first attempt at a regular expression which matches
these lines was:
>            
'<space>*,?KEY_COLUMN<space>*(VAR)?CHAR([0-9]*)'

In sed, you need (.....) instead of (.....) to group
multiple
characters, and use ? instead of ? to match zero/one
preceding
char/chars_group. try the pattern:

 
<space>*,?KEY_COLUMN<space>*(VAR)?CHAR([0-9]
*)

for example:
bash ~$ echo '
         CREATE TABLE db.table {
         KEY_COLUMN           VARCHAR(20)
        ,KEY_COLUMN_02        CHAR(20)
        ,KEY_COLUMN_03        TIMESTAMP(6)
        ,NOT_A_KEY_COLUMN     TIMESTAMP(6)
        ,CREATE_MODULE        VARCHAR(1024)
        ..........
};
' | sed -n '/ *,?KEY_COLUMN *(VAR)?CHAR([0-9]*)/p'
         KEY_COLUMN           VARCHAR(20)

> The above expression does not work. What does work is:
>            
'<space>*[<space>,]KEY_COLUMN<space>*[A-Z]
*CHAR[0-9]*)'
>
> The first one makes a lot more sense to me:
>             '<space>*[<space>,]'
> seems like a confusing way of saying 'any number of
spaces followed by
> an optional comma', which is what I thought
'<space>*,?' meant
> and
>
>             '[A-Z]*CHAR'
> does what I need it to do, but I still don't understand
what's wrong
> with '(VAR)?CHAR' As I said, I have an expression that
works, but I
> don't feel like I learned anything - I want to figure
out why what I
> had didn't work. Both failures did use question marks.
Any advice?

you may need to escape the question mark to make it a
metacharacter.

Good luck, 
Xicheng


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the
Google Groups "Regex" group.
To post to this group, send email to regexgooglegroups.com
To unsubscribe from this group, send email to
regex-unsubscribegooglegroups.com
For more options, visit this group at http://grou
ps-beta.google.com/group/regex
-~----------~----~----~----~------~----~------~--~---