python detect tab char

Question

I was trying to split words and ints inside a specific file. File's strings are in these form (line that contains word has not '\t' char but int numbers(all positive) have): (some words are numbers containing '-' char, )

-1234
\t22
\t44
\t46
absv
\t1
\t2
\t4
... 

So my idea was to split words and strings by casting the line's object to float.

def is_number(s):
    try:
        float(s)
        return True
    except ValueError:
        return False

with codecs.open("/media/New Volume/3rd_step.txt", 'Ur') as file:#open file
    for line in file: # read line by line
        temp_buffer = line.split() # split elements
        for word in temp_buffer:
            if not('-' in word or not is_number(word)):
            ....

So if it was a word I would get exceptions if not then it is a number. The file is 50 Gb , and somewhere in the middle it seems that something goes wrong with the file's format. So the only possible way to split words and numbers is with the \t char. But how can I detect it? I mean I split the line to get the string and I lose the special chars that way.

EDIT:

I am really silly and newbe sorry for wasting your time. It seems that I can find it easier with this way:

with codecs.open("/media/D60A6CE00A6CBEDD/InvertedIndex/1.txt", 'Ur') as file:#open file
    for line in file: # read line by line
    if not '\t' in line:
            print line

Solution

You should try specifying your argument to split() instead of just using the default, which is all whitespace characters. You could have it initially split on all whitespace except \t. Try this:

white_str = list(string.whitespace)    # string.whitespace contains all whitespace.
white_str.remove("\t")                 # Remove \t
white_str = ''.join(white_str)         # New whitespace string, without \t

Then instead of split(), use split(white_str). This will split your lines on all whitespace except for \t to get your strings. Then you can detect \t later on for what you need.