BeautifulSoup and Problem Tables

[academicedgar]

Hi

I would appreciate some help. I am trying to learn Python and want to
use BeautifulSoup to pull some data from tables. I was really psyched
earlier tonight when I discovered that I could do this

from BeautifulSoup import BeautifulSoup
bst=file(r"c:\bstest.htm").read()
soup=BeautifulSoup(bst)
rows=soup.findAll('tr')
len(rows)
a=len(rows[0].findAll('td'))
b=len(rows[1].findAll('td'))
c=len(rows[2].findAll('td'))
d=len(rows[3].findAll('td'))
e=len(rows[4].findAll('td'))
f=len(rows[5].findAll('td'))
g=len(rows[6].findAll('td'))
h=len(rows[8].findAll('td'))
i=len(rows[9].findAll('td'))
j=len(rows[10].findAll('td'))
k=rows[1].findAll('td')[1].contents[0]
So here I am chortling to myself thinking this is too easy. I know
that the data columns are in rows[0] and so I can learn some more
python to figure out how to create tuples so I can lable each data
item using the row and column headings plucked from the contents.

However, I discovered that my tables have inconsistent numbers of
rows. Even though the tables look pretty. It might be that the
column heading for the third column is "Apples" but the value for
"Apples" in the fourth row is not in the third position in the row but
the fourth.

Now I am reluctant to make any assumptions because the tables were
created inconsistently. What I mean is that in some tables if there is
no value for a row/column intersection then there is a blank line, in
other tables if there is no value for a row/column intersection then
the length of k (as above) is 0.

I have been Googling for some insight into this and I have not been
successful finding anything. I would really appreciate any suggestions
or some direction about how to better describe the problem.

[Peter Pearson]

On Sat, 20 Sep 2008 20:51:52 -0700 (PDT), ac***********@gmail.com wrote:
[snip]

from BeautifulSoup import BeautifulSoup
bst=file(r"c:\bstest.htm").read()
soup=BeautifulSoup(bst)
rows=soup.findAll('tr')
len(rows)
a=len(rows[0].findAll('td'))
b=len(rows[1].findAll('td'))
c=len(rows[2].findAll('td'))
d=len(rows[3].findAll('td'))
e=len(rows[4].findAll('td'))
f=len(rows[5].findAll('td'))
g=len(rows[6].findAll('td'))
h=len(rows[8].findAll('td'))
i=len(rows[9].findAll('td'))
j=len(rows[10].findAll('td'))
k=rows[1].findAll('td')[1].contents[0]

[snip]

However, I discovered that my tables have inconsistent numbers of
rows.

[snip]

I have been Googling for some insight into this and I have not been
successful finding anything. I would really appreciate any suggestions
or some direction about how to better describe the problem.

Would it be accurate to describe the problem as wanting to
extract the contents of the cth column of the rth row of a
table in spite of various pathologies in the construction of
the table?

If so, maybe it would help to post sample HTML (trimmed to a
minimum) of the pathologies that must be handled. I gotta
confess, though, that it doesn't take many rowspans or colspans
to put this problem beyond my reach.

--
To email me, substitute nowhere->spamcop, invalid->net.

[clurks]

ac***********@gmail.com wrote:

Hi

I would appreciate some help. I am trying to learn Python and want to
use BeautifulSoup to pull some data from tables. I was really psyched
earlier tonight when I discovered that I could do this

from BeautifulSoup import BeautifulSoup
bst=file(r"c:\bstest.htm").read()
soup=BeautifulSoup(bst)
rows=soup.findAll('tr')
len(rows)
a=len(rows[0].findAll('td'))
b=len(rows[1].findAll('td'))
c=len(rows[2].findAll('td'))
d=len(rows[3].findAll('td'))
e=len(rows[4].findAll('td'))
f=len(rows[5].findAll('td'))
g=len(rows[6].findAll('td'))
h=len(rows[8].findAll('td'))
i=len(rows[9].findAll('td'))
j=len(rows[10].findAll('td'))
k=rows[1].findAll('td')[1].contents[0]
So here I am chortling to myself thinking this is too easy. I know
that the data columns are in rows[0] and so I can learn some more
python to figure out how to create tuples so I can lable each data
item using the row and column headings plucked from the contents.

However, I discovered that my tables have inconsistent numbers of
rows. Even though the tables look pretty. It might be that the
column heading for the third column is "Apples" but the value for
"Apples" in the fourth row is not in the third position in the row but
the fourth.

Now I am reluctant to make any assumptions because the tables were
created inconsistently. What I mean is that in some tables if there is
no value for a row/column intersection then there is a blank line, in
other tables if there is no value for a row/column intersection then
the length of k (as above) is 0.

I have been Googling for some insight into this and I have not been
successful finding anything. I would really appreciate any suggestions
or some direction about how to better describe the problem.

This may help, and it may not.

One of the most useful features of BeautifulSoup is the .prettify()
function. Use it, and you can examine your xml with a browser or
decent editor. You will be able, hopefully, to see attributes that
may explain your difficulties. The one that snuck up and bit me was
table:number-columns-repeated -- I don't know what is biting you,
but it is weird the way known cells are completely skipped with this
attribute -- you have to know about it and check for it.

If there is a way to deal with that attribute using BeautifulSoup I
was unable to find it given the documentation and examples available
to me. I took a step back and tried xml.sax -- it is straighforward
and easy to use -- you build an xml.sax.handler for your xml, and it
deals with it.

It worked for me, but, as always, YMMV

sc