newb: BeautifulSoup

[jobs]

I am trying to use BeautifulSoup:
soup = BeautifulSoup(page)

td_tags = soup.findAll('td')
i=0
for td in td_tags:
i = i+1
print "td: ", td
# re.search('[0-9]*\.[0-9]*', td)
price = re.compile('[0-9]*\.[0-9]*').search(td)

I am getting an error:

price= re.compile('[0-9]*\.[0-9]*').search(td)
TypeError: expected string or buffer

[jobs]

soup = BeautifulSoup(page) 
 
    td_tags = soup.findAll('td') 
     i=0 
     for td in td_tags: 
             i = i+1 
             print "td: ", td 
         #  re.search('[0-9]*\.[0-9]*', td) 
             price = re.compile('[0-9]*\.[0-9]*').search(td)

I am getting an error:

price= re.compile('[0-9]*\.[0-9]*').search(td)
TypeError: expected string or buffer

td is an object. I need to pass a string into re.search. I need to pass a string to re.compile().search(). Anyone how I can convert object to string?
Anyone know how to fix this?

[BearofNH]

Thank you for reposting via [code] tags, and using an appropriate title.

td is an instance, not a string.

You probably want to say price = re.compile('[0-9]*\.[0-9]*').search(td.contents[0]), and/or iterate over td.contents[] (via [inlinecode]for tdc in td.contents:, say).

You know, of course, that you can write something like mypat = re.compile('[0-9]*\.[0-9]*') and then later just do mypat.search(td.contents[0]), which is usually more efficient in both memory and CPU, if you call re.compile() more than once.

[jobs]

mypat = re.compile('[0-9]*\.[0-9]*')

This pattern only works if it is floating point. I need construct it so it works for both decimal and whole numbers.

I have scenario where I could get either decimal numbers(1234.890) or whole numbers(10000).

How do you get the proper regex?

[BearofNH]

There's lots of ways, depending on how paranoid you need to be. One way, you make the decimal and following optional, as in: mypat = re.compile('[0-9]*(\.[0-9]*|$)') , with or without the $ depending on your needs. Regexp matching will try to grab as much as it can, which does what you want here.