Re: Logistic Regression Match Error

2016-11-19 Thread Meeraj Kunnumpurath
Thank you, it was the escape character, option("escape", "\"") Regards On Sat, Nov 19, 2016 at 11:10 PM, Meeraj Kunnumpurath < mee...@servicesymphony.com> wrote: > I triied .option("quote", "\""), which I believe is the default, still the > same error. This is the offending record. > > Primo 4-I

Re: Logistic Regression Match Error

2016-11-19 Thread Meeraj Kunnumpurath
I triied .option("quote", "\""), which I believe is the default, still the same error. This is the offending record. Primo 4-In-1 Soft Seat Toilet Trainer and Step Stool White with Pastel Blue Seat,"I chose this potty for my son because of the good reviews. I do not like it. I'm honestly baffled b

Re: Logistic Regression Match Error

2016-11-19 Thread Meeraj Kunnumpurath
Digging through it looks like an issue with reading CSV. Some of the data have embedded commas in them, these fields are rightly quoted. However, the CSV reader seems to be getting to a pickle, when the records contain quoted and unquoted data. Fields are only quoted, when there are commas within t

Logistic Regression Match Error

2016-11-19 Thread Meeraj Kunnumpurath
Hello, I have the following code that trains a mapping of review text to ratings. I use a tokenizer to get all the words from the review, and use a count vectorizer to get all the words. However, when I train the classifier I get a match error. Any pointers will be very helpful. The code is below