Pythonic way to conditionally iterate over items in a list

New to programming in general, so I'm probably going about this the wrong way. I'm writing an lxml parser where I want to omit HTML table rows that have no content from the parser output. This is what I've got:

for row in doc.cssselect('tr'):
    for cell in row.csssele开发者_StackOverflow社区ct('td'):
        sys.stdout.write(cell.text_content() + '\t')
    sys.stdout.write '\n'

The write() stuff is temporary. What I want is for the loop to only return rows where tr.text_content != ''. So I guess I'm asking how to write what my brain thinks should be 'for a in b if a != x' but that doesn't work.

Thanks!

for row in doc.cssselect('tr'):
    cells = [ cell.text_content() for cell in row.cssselect('td') ]
    if any(cells):
        sys.stdout.write('\t'.join(cells) + '\n')

prints the line only if there is at least one cell with text content.

ReEdit:

You know, I really don't like my answer at all. I voted up the other answer but I liked his original answer because not only was it clean but self explanatory without getting "fancy" which is what I fell victim to:

for row in doc.cssselect('tr'):
    for cell in row.cssselect('td'):
        if(cel.text_content() != ''):
            #do stuff here

there's not much more of an elegant solution.

Original-ish:

You can transform the second for loop as follows:

[cell for cell in row.cssselect if cell.text_content() != '']

and turn it into a list-comprehension. That way you've got a prescreened list. You can take that even farther by looking at the following example:

a = [[1,2],[2,3],[3,4]
newList = [y for x in a for y in x]

which transforms it into [1, 2, 2, 3, 3, 4]. Then you can add in the if statement at the end to screen out values. Hence, you'd reduce that into a single line.

Then again, if you were to look at itertools:

ifilter(lambda x: x.text_content() != '', row.cssselect('td'))

produces an iterator which you can iterate over, skipping all items you don't want.

Edit:

And before I get more downvotes, if you're using python 3.0, filter works the same way. No need to import ifilter.