preg_match_all optional subpattern

Using preg_match_all, I need to capture a list of first and last names plus
an optional country code proceeding them.

For example:

<tr><td>AU</td><td>Jane Smith</td></tr>
<tr><td></td><td>Bill Johnson</td></tr>
<tr><td>GB</td><td>Larry Brown</td></tr>
<tr><td>US</td><td>Mary Jordon</td></tr>
<tr><td></td><td>Peter Jones</td></tr>

The country code might exist, it might not.

I would like the array contents to look like this:

AU Jane Smith
Bill Johnson
US Larry Brown
GB Mary Jordon
Peter Jones

I know a subpattern is needed of all the possible country codes:

AU|GB|US

but how do you include this as an optional subpattern?

Thanks in advance.

Han wrote:

> Using preg_match_all, I need to capture a list of first and last names plus
> an optional country code proceeding them.
>
> For example:
>
> <tr><td>AU</td><td>Jane Smith</td></tr>
> <tr><td></td><td>Bill Johnson</td></tr>
> <tr><td>GB</td><td>Larry Brown</td></tr>
> <tr><td>US</td><td>Mary Jordon</td></tr>
> <tr><td></td><td>Peter Jones</td></tr>
>
> [...] I know a subpattern is needed of all the possible country codes:
>
> AU|GB|US
>
> but how do you include this as an optional subpattern?

The ? quantifier means zero or one of whatever came before,
representable by {0,1}. Quantifying a subpattern using the
question mark denotes its nonobligatory nature.

So, to match optional two-letter country codes within a table cell
(doesn't properly cater for attributes, but that's rectifiable):

`<td.*>([a-z]{2})?</td.*>`Usi

If you wish to list the possible values, precluding others:

`<td.*>(au|gb|us)?</td.*>`Usi

--
Jock

"Han" <nobody@nowhere.com> wrote in message news:<TT4gb.225533$mp.141550@rwcrnsc51.ops.asp.att .net>...
> Using preg_match_all, I need to capture a list of first and last names plus
> an optional country code proceeding them.
>
> For example:
>
> <tr><td>AU</td><td>Jane Smith</td></tr>
> <tr><td></td><td>Bill Johnson</td></tr>
> <tr><td>GB</td><td>Larry Brown</td></tr>
> <tr><td>US</td><td>Mary Jordon</td></tr>
> <tr><td></td><td>Peter Jones</td></tr>
>
> The country code might exist, it might not.
>
> I would like the array contents to look like this:
>
> AU Jane Smith
> Bill Johnson
> US Larry Brown
> GB Mary Jordon
> Peter Jones
>
> I know a subpattern is needed of all the possible country codes:
>
> AU|GB|US

this pattern should do the job:

"{<tr><td>\s*([A-Z]{2})?\s*</td><td>\s*(\w+)?\s*(\w+)\s*</td></tr>}im"

if this pattern is used in preg_match_all, it should produce the
desired result.
it will extract the country code if availible, first name if
availible, and last name.
They will be put in an 2 dim array. If no country code or first name
is given, the array element will be left empty.

hope this helps,

sascha



>
> but how do you include this as an optional subpattern?
>
> Thanks in advance.

John,

Thank you for another detailed reply.

The cryptic syntax is beginning to slowly sink in, but there's still a few
nagging issues.

In my price list, the amount may or may not be preceded with a $ sign.

For instance, the list might look like this:

$2.99
1.99
$3.00
$4.00

I modified my price pattern to accommodate this:

((\\$|\s*)?\d{1,3}\.\d{2})

which works great. The problem is, it also creates another array dimension
that contains only $ or space:

$

$
$

I can simply ignore this dimension, but is there a way to prevent it?

Thanks (again) in advance.

"John Dunlop" <john+usenet@johndunlop.info> wrote in message
news:MPG.19eb262ef3b81717989777@news.freeserve.net ...
> Han wrote:
>
> > Using preg_match_all, I need to capture a list of first and last names
plus
> > an optional country code proceeding them.
> >
> > For example:
> >
> > <tr><td>AU</td><td>Jane Smith</td></tr>
> > <tr><td></td><td>Bill Johnson</td></tr>
> > <tr><td>GB</td><td>Larry Brown</td></tr>
> > <tr><td>US</td><td>Mary Jordon</td></tr>
> > <tr><td></td><td>Peter Jones</td></tr>
> >
> > [...] I know a subpattern is needed of all the possible country codes:
> >
> > AU|GB|US
> >
> > but how do you include this as an optional subpattern?
>
> The ? quantifier means zero or one of whatever came before,
> representable by {0,1}. Quantifying a subpattern using the
> question mark denotes its nonobligatory nature.
>
> So, to match optional two-letter country codes within a table cell
> (doesn't properly cater for attributes, but that's rectifiable):
>
> `<td.*>([a-z]{2})?</td.*>`Usi
>
> If you wish to list the possible values, precluding others:
>
> `<td.*>(au|gb|us)?</td.*>`Usi
>
> --
> Jock

Han wrote:

> ((\\$|\s*)?\d{1,3}\.\d{2})
>
> which works great. The problem is, it also creates another array
> dimension that contains only $ or space:
>
> [...] I can simply ignore this dimension, but is there a way to
> prevent it?

Subpatterns that begin with the two character sequence "?:" aren't
captured. You could then write your pattern as:

`(?:\\$|\s*)?\d{1,3}\.\d{2}`

--
Jock

Jock,

That's it--thanks.

I've been spending some time re-reading the pattern documentation on php.net
and it's beginning to sink in.

Again, much appreciated!

"John Dunlop" <john+usenet@johndunlop.info> wrote in message
news:MPG.19ec77cdbce17be098977c@news.freeserve.net ...
> Han wrote:
>
> > ((\\$|\s*)?\d{1,3}\.\d{2})
> >
> > which works great. The problem is, it also creates another array
> > dimension that contains only $ or space:
> >
> > [...] I can simply ignore this dimension, but is there a way to
> > prevent it?
>
> Subpatterns that begin with the two character sequence "?:" aren't
> captured. You could then write your pattern as:
>
> `(?:\\$|\s*)?\d{1,3}\.\d{2}`
>
> --
> Jock