[Developers] Long requests in python with metaweb.readall

[Scott Meyer]

Alexei Ustyuzhaninov wrote:
> Hello,
> 
> I have this request which is expected to return several millions records:
> 
> {
>   "album" : [
>     {
>       "artist" : []
>     }
>   ],
>   "artist" : null,
>   "id" : null,
>   "name" : null,
>   "type" : "/music/track"
> }
> 
> When I run this request with metaweb.readall function it hangs forever
> until an external exception (such as timeout) will happen. I modified
> readall a little (see metaweb2.py attached) in order to check the
> uniqueness of the intermediate data obtained from Freebase. And yes, it
> discovered that the data start duplicated after fetching 27100 first
> records.
> 
> Can somebody help with this problem?

Yes.  This is a known problem. We plan on deploying a fix in the
next week, or so.

If all you want is musical tracks, (or instances of any other type) you
can download those as files:

http://download.freebase.com/datadumps/2009-01-13/browse/music/track.tsv

-Scott