Edit: Solved - see Ian's comment below

I'm not a dropbox user, but if you copy your link into a browser it displays a page that asks for a login/ignore. Is there a public link that dropbox provides so that the link when entered into a browser just displays the file (without the login/ignore page)?

Edit: Make sure you are not currently logged into dropbox to test this.

Thanks Rob,

I get the same thing with Firefox, but I can bypass the link asking me to login, and subsequently download the file. I just pasted the link into my Chrome browser (after logging out of dropbox), and it does ask me to login, but if I bypass this, I can download.

Maybe there is an option in dropbox to avoid the login request? Maybe this is getting in the way?

Kinda confusing, since the Quantopian error makes it sound like the file is accessible, but just can't be parsed.

Grant

Grant,

Try this url and see if it works. (I just copied your file to a test server of mine).

Edit: URL removed

Works! So I wonder what the trick is with dropbox? Or I could use an alternative service? Any recommendations? Thanks, Grant

Great. You could use Google Sheets (free with a Google account) and output as a CSV file. Just click "Share" and generate a public link (viewable by anyone without a sign-in) and then you have to tweak the url it gives you to generate CSV output. Try this link I just made of your file (and note the "/export?format=csv" addition):

https://docs.google.com/spreadsheets/d/17jKqzzDfpmPb7TlXMuvc7aWOspIuBdzw_35f2OGCE6I/export?format=csv

Hey Grant,

You can replace the 'www.dropbox.com' domain portion of any dropbox share URL with the domain 'dl.dropboxusercontent.com', i.e.,

https://dl.dropboxusercontent.com/s/i1teg8xcokks6fl/fetcher_test.csv?dl=0

Cheers,
Ian

Ian - much easier!

You may still want to consider using Google Sheets as a handy way to manually enter data or automatically integrate multiple data sources into one fetcher file since it allows for XML/HTML/CSV imports from other sources. I'm considering doing this based on Grant's suggestion here to "remote control" a running algorithm.

I use a public folder in Dropbox which means I can right-click any file and choose "copy public link". It really is extremely convenient. You have to be a paying user for that feature though.

Thanks all for the help. Another question. The code below results in the error:

24 Error Runtime exception: AttributeError: 'SIDData' object has no attribute 'custom'

It will run if I use the standard attribute, 'price', but I'd like to use some custom ones. Is this possible? Or is there an alternative approach?

Grant

import pandas

def rename_col(df):  
    # df = df.rename(columns={'my_data': 'price'})  
    df = df.rename(columns={'my_data': 'custom'})  
    return df

def initialize(context):  
    # fetch_csv('https://www.dropbox.com/s/i1teg8xcokks6fl/fetcher_test.csv?dl=0',  
    #            date_column = 'day',  
    #            date_format = '%m/%d/%y',  
    #            symbol = 'my_data')  
    fetch_csv('http://205.186.154.106/q/fetcher_test.csv',  
               date_column = 'day',  
               date_format = '%m/%d/%y',  
               symbol = 'my_data',  
               post_func=rename_col)  
    context.spy = sid(8554)  
def handle_data(context, data):  
    # print data['my_data'].price  
    print data['my_data'].custom  

Drop box has terminated the public link. I try to use the google sheet, but it does not work either. I get the following error:
AttributeError: 'DataFrame' object has no attribute 'Date'

This is my link:
https://docs.google.com/spreadsheets/d/13-EFqu0D-8Q9L_9fjWaf8vr2SWgFm0_vgUeQqWarZg0/export?format=csv

Can someone help me with this?