We help IT Professionals succeed at work.

Python TypeError

GVNPublic123
GVNPublic123 asked
on
Ok, so these lines usually work fine, but sometime produce type error (the bold part is producing it).
 1228         url = "http://www.google.com/" % params
 1229        [b] output = getPage("https://api-ssl.bitly.com/v3/shorten?login=xxx&apiKey=xxx&%s&format=txt"\
 1230         % urllib.urlencode({'longUrl':url}))[:22].replace("http://", "")[/b]

Open in new window


Error:
TypeError: unsubscriptable object
      args = ('unsubscriptable object',)

Also in debug report, when this error occured both params and url variables were setup fine. HELP!
Comment
Watch Question

CERTIFIED EXPERT

Commented:
You should copy/paste our error message with the trace.

Author

Commented:
But I gave you traced line of code and error...what more do you need? There aint no more in report.
CERTIFIED EXPERT

Commented:
How do you run it?  Is it a part of some web server code?  Where the 'args = ...' came from?

Author

Commented:
No idea...its just a random python fail.

Author

Commented:
It occured, but I cant recreate it. I put exception handling for next time.
CERTIFIED EXPERT

Commented:
How do you run your Python code?  Is it saved in a file and launched say from command line?  How do you know that the mentioned line caused the error?  What version of Python do you use?  What OS do you use?

Author

Commented:
CentOS 5, its a website, python 2.4.3

Got it from bug report.
CERTIFIED EXPERT
Commented:
OK.  I cannot help too much here.  If you know it was the line
1229  output = getpage(...)[:22].replace("http://", "")[/b]

Open in new window


The error generally means that you try to index/slice what cannot be indexed -- i.e. the object you probably got cannot be sliced using the [:22].  It is likely that the getpage() returned None as the error result or something like that.

Try to avoid that kind of method chaining if the code is a bit complex and if it can fail:
    output = ''         # or some better initial value
    page = getpage(...)
    if page:              # or some more explicit testing
        output = page[:22].replace(...)

Open in new window

Author

Commented:
Yeh, I put exception handling in so now I dont care if slicing fails. The worker will pick it up later automatically if it fails on request.
CERTIFIED EXPERT

Commented:
Sorry I could not be more explicit.  Thanks for the points, and have a good time ;)