> ValueError: IDs are not in the same order in each feature set

I'm getting this error when running e.g. `join_features all_pos/*/train/*jsonlines all_pos.jsonlines`. I assume I can write a small script to fix this.

Posted by: ndronen @ Jan. 18, 2018, 7:13 p.m.

By doing "`join_features all_pos/*/train/*jsonlines all_pos.jsonlines", you are doing a join across all genres. This is not possible because of the different texts and tokens of each genre. You can do a successful join by going into a specific genre e.g. "academic" and do the following:

join_features train/*jsonlines test.jsonlines
Loading train/C-BiasDown.jsonlines... done
Loading train/C-BiasUp.jsonlines... done
Loading train/CCDB-BiasUpDown.jsonlines... done
Loading train/Corpus.jsonlines... done
Loading train/P.jsonlines... done
Loading train/T.jsonlines... done
Loading train/U.jsonlines... done
Loading train/UL.jsonlines... done
Loading train/VN-Raw.jsonlines... done
Loading train/WordNet.jsonlines... done
Writing test.jsonlines...done

Posted by: cleong @ Jan. 22, 2018, 7:05 p.m.
Post in this thread