Uploading CSV and excel files

Using Streamlit
1

Hi,

Thanks for your great work!

My website allows the user to upload CSV files. I would like to add the possibility, using the same widget, to upload Excel files. It does not work I think for an issue of encoding that I was unable to solve.

At the moment I upload csv files like so:

    try:
          bytesData = uploadedFile.getvalue()
          encoding = encodingUTF8 
          s=str(bytesData,encoding)
          result = StringIO(s) 
    except:
          bytesData = uploadedFile.getvalue()
          encoding = encodingISO 
          s=str(bytesData,encoding)
          result = StringIO(s)

This works fine. I wanted to add the possibility for the user also to upload xlsx files, so I changed the file uploaded code to

        uploadedFile = st.file_uploader(fileUploadLabel, type=['csv','xlsx'],accept_multiple_files=False,key="fileUploader")

and added a “read excel” bit

  try:
        df=pd.read_csv(dataUploaded, error_bad_lines=True, warn_bad_lines=False,sep=chosenFileSeparator)
  except:
        try:
              df = pd.read_excel(dataUploaded)
        except:      
              df=pd.DataFrame()

I get this error (if I take out the “try” or course)

image
image653Ă—633 20.2 KB

The identical file just in CSV instead of Excel gets read with not problem.

Thanks so much for your help!

Fabio

1 Like
2

@Fabio Hi, you can just use st.file_uploader(“the notice you want to tell”) to finish upload function.
both csv and xls or xlsx file can be uploaded.
example:

import pandas as pd
import streamlit as st
uploaded_file = st.file_uploader(“Choose a file”)
if uploaded_file is not None:
#read csv
df1=pd.read_csv(uploaded_file)

   #read xls or xlsx
   df1=pd.read_excel(uploaded_file)

else:
st.warning(“you need to upload a csv or excel file.”)

if you want to capture the file name of uploaded, you can write code like this:
filename=uploaded_file.name

1 Like
3

Hi @BeyondMyself,

thanks for your answer!

In my case the problem is that I do not know beforehand whether the file will be CSV or xlsx.

Since (I suppose) csv and excel files have a different encoding, I suppose that something bad happens in my encoding function

   try:
      bytesData = uploadedFile.getvalue()
      encoding = encodingUTF8 
      s=str(bytesData,encoding)
      result = StringIO(s)

I do not get an error here, but then when I run “read_excel” on the object returned from my “StringIO(s)” function, I get the error.

try:
          df = pd.read_excel(dataUploaded)
    except:      
          df=pd.DataFrame()

Are you saying that filename=uploaded_file.name will give me also the file suffix (csv or xlsx) :slight_smile: not just the file name ?

Thanks again!

Fabio

4

Yes it returns also the suffix!!! :slight_smile: