Easy Regular Expressions Question

[William Wagers]

I have this pattern:
<b>(.+?)</b>.*?<td valign=top>(.+?)<br>(.*?)<br><small>.*?<td valign=top><nobr>(.+?)</nobr>
but I need this part:
<br>(.*?)
to be optional. In other words, that whole piece may be missing. For example, I could encounter 1<br>2 or just 1 at that point.
How should I code that part?
Thanks.

[Jim Yingst]

Do you need to capture the taxt after that optional <br> if it's there? If not, the simplest solution is just to ignore that one <rb> entirely, and just look for the <br><small> part:
<b>(.+?)</b>.*?<td valign=top>(.+?)<br><small>.*?<td valign=top><nobr>(.+?)</nobr>
Even if you want to know the details of whether the <br> is there or not - you can test group 2 afterwards, in a separate regex. You don't have to do everything in one expression - it can be to confusing.
Here's another alternative - I inserted whitespace and comments for clarity; compile with Pattern.COMMENTS so these will be ignored.

<b>(.+?)</b>.*?<td valign=top>
(?:
    (.+?)<br>(.*?)   # look for the <br> first
|
    (.+?)            # then try with no <br>
)
<br><small>.*?<td valign=top><nobr>(.+?)</nobr>

I think this will lead to a lot of backtracking which may impact performance. Better may be the following, using negative lookahead:

<b>(.+?)</b>.*?<td valign=top>
(
    (?:              
        (?!<br>)  .    # anything as long as it's not the start of a <br>
    )+               # as many times as possible
)
(<br>)  ?
(.*?)
<br><small>.*?<td valign=top><nobr>(.+?)</nobr>

[ August 05, 2003: Message edited by: Jim Yingst ]

[Stefan Wagner]

can't you nest braces like this:
(<br>(.*?))?
?
[ August 05, 2003: Message edited by: Stefan Wagner ]

[Jim Yingst]

can't you nest braces like this:
(<br>(.*?))?
Yeah. When I first considered that possibility I was thinking that this would not interact well (efficiency-wise) with the (.+?) immediately before it. I distrust having too many *'s, +'s and ? nested in close proximity, as they often do strange things. But as I think more about this specific example I guess it's OK.

<b>(.+?)</b>.*?<td valign=top>
(.+?)(<br>(.*?))?
<br><small>.*?<td valign=top><nobr>(.+?)</nobr>

I still like my very first answer best though.
[ August 05, 2003: Message edited by: Jim Yingst ]

[William Wagers]

Thanks for the suggestions, but neither one works for me. Both solutions, leave a <br> in the data, and/or get lost, presumably because of the <br><small> which follows.
I was a little confused by the extra group(s). I was able to make it work for me. Thanks.
[ August 08, 2003: Message edited by: William Wagers ]