CODEWITHSUNDEEP
pythonpython-2.7xlrd
francisco sollima
francisco sollima

Reputation: 8338

Failing to open an Excel file with Python

I'm on a Debian GNU/Linux computer, working with Python 2.7.9.

As a part of my job, I have been making python scripts that read inputs in various formats (e.g. Excel, Csv, Txt) and parse the information to more standarized files. It's not my first time opening or working with Excel files.

There's a particular file which is giving me problems, I just can't open it. When I tried with xlrd (version 0.9.3), it gave me the following error:

xlrd.open_workbook('sample.xls')

XLRDError: Unsupported format, or corrupt file: BOF not workbook/worksheet: op=0x0009 vers=0x0002 strm=0x000a build=0 year=0 -> BIFF21

I tried to investigate the matter on my own, found a couple of answers in StackOverflow but I couldn't open it anyway. This particular answer I found may be the problem (the second explanation), but it doesn't include a workaround: https://stackoverflow.com/a/16518707/4345659

A tool that could conert the file to csv/txt would also solve the problem.

I already tried with:

A sample file is available here: https://ufile.io/r4m6j

As a side note, I can open it with LibreOffice Calc and MS Excel, so I could eventually change it to csv that way. The thing is, I need to do it all with a python script.

Thanks in advance!

Upvotes: 5

Views: 12004

Answers (9)

Shmygol
Shmygol

Reputation: 933

It seems like MS Problem. The xls file is very strange, maybe you should contact xlrd support.

But I have a crazy workaround for you: xls2ods. It works for me even though xls2csv doesn't (SiC!).

So, install catdoc first:

$sudo apt-get install catdoc

Then convert your xls file to ods and open ods using pyexcel_ods or whatever you prefer. To use pyexcel_ods install it first using pip install pyexcel_ods.

import subprocess

from pyexcel_ods import get_data

file_basename = 'sample'

returncode = subprocess.call(['xls2ods', '{}.xls'.format(file_basename)])
if returnecode > 0:
    # consider to use subprocess.Popen if you need more control on stderr
    exit(returncode)

data = get_data('{}.ods'.format(file_basename))
print(data)

I'm getting following output:

OrderedDict([(u'sample',
              [[u'labo',
                u'codfarm',
                u'farmacia',
                u'direccion',
                u'localidad',
                u'nom_medico',
                u'matricula',
                u'troquel',
                u'producto',
                u'cant_total']])])

Upvotes: 2

Bruce Z. Wang
Bruce Z. Wang

Reputation: 1

Unable to open your Excel either. Just as yadayada said, I think it is the problem of data source. If you really want to figure out the reason, I suggest you ask questions about the excel instead of python.

Upvotes: 0

snake_charmer
snake_charmer

Reputation: 3013

Here is a kludge I would use:

Assuming you have LibreOffice on Debian, you could either convert all your *.xls files into *.csv using:

import os 
os.system("libreoffice --headless --convert-to csv *.xls")
#or use os.call

... and then work consistently with csv.

Or you could convert only the corrupted file(s) when needed using a try/except block:

import os 

try:
    xlrd.open_workbook('sample.xls')
except XLRDError:
    os.system("libreoffice --headless --convert-to csv sample.xls")
    # mycsv = open("sample.csv", "r")
    # for line in mycsv.readlines():
    #     ...
    #     ...

OBS: Keep LibreOffice closed while running the script.

Alternatively there are other tools out there to do the conversion. Here is one (which I have not tested): https://github.com/dilshod/xlsx2csv

Upvotes: 2

yadayada
yadayada

Reputation: 357

The file you provided is corrupted, so there is no way for other responders to test it and recommend a good solution. And exception you posted confirming that. As a solution you can try to debug some things, please see some steps below:

  • You mentioned you tried the xlrd library. Try to check if your xlrd module is upto date by executing this:

    Python 2.7.9

    >>> import xlrd

>>> xlrd.__VERSION

update to the latest official version if needed

  • Try to open any other *.xls file and see if it works with Python version you're using and current library.

  • Check module documentation it's pretty good, and there are some different things described how to use this module on various platforms( Win vs. Linux)http://xlrd.readthedocs.io/en/latest/dates.html

  • You always can rich out to the community (there is still a chance that you might be getting into some weird state or bug) the link is here https://github.com/python-excel/xlrd/issues

Hope that helps.

Upvotes: 0

ChaosPredictor
ChaosPredictor

Reputation: 4051

If you need to get the words as a list of strings:

text_file = open("sample.xls", "r")
lines = text_file.read().replace(chr(200), '').replace(chr(0), '').replace(chr(1), '').replace(chr(5), '').replace(chr(2), '').replace(chr(3), '').replace(chr(4), '').replace(chr(6), '').replace(chr(7), '').replace(chr(8), '').replace(chr(9), '').replace(chr(10), '').replace(chr(12), '').replace(chr(15), '').replace(chr(16), '').replace(chr(17), '').replace(chr(18), '').replace(chr(49), '').replace('Arial', '')
for line in lines.split(chr(128)):
    print(line)

the output: enter image description here

Upvotes: 0

ChaosPredictor
ChaosPredictor

Reputation: 4051

It's a file format issue. I'm not sure what file type is it but it's not Excel. I just open and saved the file with sample2.xls name and compare the types: enter image description here

How are you creating this file?

Upvotes: 0

user5852223
user5852223

Reputation:

This code open an xls file and saves it as a cvs file, using the comtypes package:

import comtypes.client as cl
progId = "Excel.Application.15"
xl = cl.CreateObject(progId)
wb = xl.Workbooks.Open(r"C:\Users\aUser\Desktop\thermoList.xls")
wb.SaveAs(r"C:\Users\aUser\Desktop\thermoList.csv",FileFormat=6)
xl.DisplayAlerts = False
xl.Quit()

I could not test it with "sample.xls" which is corrupt. Your could try with another file. You might need to adjust the progId according to your version of Excel.

Upvotes: 0

user5852223
user5852223

Reputation:

If you are targeting windows, if you have Excel installed, and if you are familiar with Excel VBA, you will have a quick solution using the comtypes package:

http://pythonhosted.org/comtypes/

You will have direct access to Excel by its COM interfaces.

Upvotes: 0

Vasily Bronsky
Vasily Bronsky

Reputation: 433

It's always work for me with any xls or xlsx files:

def csv_from_excel(filename_xls, filename_csv):
    wb = xlrd.open_workbook(filename_xls, encoding_override='YOUR_ENCODING_HERE (f.e. "cp1251"')
    sh = wb.sheet_by_index(0)
    your_csv_file = open(filename_csv, 'wb')
    wr = unicodecsv.writer(your_csv_file)
    for rownum in xrange(sh.nrows):
        wr.writerow(sh.row_values(rownum))
    your_csv_file.close()

So, i don't work directly with excel file before convert them to csv. Mb it will help you

Upvotes: -2

Related Questions