Fread options not extended to excel files

[samukweku]

• Did you find a bug in datatable, or maybe the bug found you?
  I cant select columns or limit number of rows read when working on an excel file

• How to reproduce the bug?

import urllib

url = 'https://github.com/samukweku/Extracting-Data-from-Excel-with-Python/blob/master/names.xlsx?raw=true'

filename = 'test.xlsx'
        
urllib.request.urlretrieve(url, filename)

fread("test.xlsx/naija")
      Name	      Age	Height
0	Tolu	        24       2
1	Chukwuka 	50	1.8
2	Ogor	        15	1.5

# column selection fails
fread('test.xlsx/naija', columns={"Name"})

# nrows fails as well
fread("test.xlsx/naija", max_nrows=2)

• What was the expected behavior?

  # column selection : 
fread('test.xlsx/naija', columns={"Name"})
      Name
0	Tolu
1	Chukwuka
2	Ogor

# row selection : 
fread("test.xlsx/naija", max_nrows=2)
      Name	     Age	Height
0	Tolu	     24       	2
1	Chukwuka     50      	1.8

• Your environment?

dt.__version__
'0.11.0a0+build.1600188275.sam'

sys.version
'3.8.5 | packaged by conda-forge | (default, Aug 29 2020, 01:22:49) \n[GCC 7.5.0]'

operating system
'Linux-5.4.0-7642-generic-x86_64-with-glibc2.10'

[pradkrish]

@samukweku I can confirm that I am able to reproduce your observation for xlsx files. I only hope it is not by design. :)

[pradkrish]

Just briefly looked into the code. I don't think it's even implemented for xls files, perhaps @st-pasha could comment on this??

[st-pasha]

That's correct. The logic for reading Excel files is completely separate from the code that reads CSV files, and consequently every option has to be implemented for Excel parsing separately.

[pradkrish]

That's what I gathered looking at the code. is there an appetite to do so for Excel files? My guess is csv files are probably much more common than Excel files in data manipulation tasks making this a low priority.

[Jiezuel]

I had the same problem when I tried to read .xlsx file.
I think the reason is probably that fread uses xlrd to read .xlsx or .xls file. See fread.py in line 325 and datatable/xls.py.
Unfortunately, xlrd updated the version on 11 Dec 2020, and it had removed support for anything other than .xls files.