Yes, We've done various fixes with regards to dataset parsing. On Mon, Jun 1, 2015 at 4:16 PM, Manorama Perera <[email protected]> wrote:
> Hi Supun, > > I've tested for quoted column headers using the linear regression sample, > by adding beginning and ending double quotes to the column headers of the > dataset. > > "fixed acidity","volatile acidity","citric acid",....., "quality" > > Model building and prediction were successful. > > As you've mentioned, the model building failure issue might have been > fixed with some earlier changes. > > Thanks. > > On Mon, Jun 1, 2015 at 2:42 PM, Supun Sethunga <[email protected]> wrote: > >> Hi Manorama, >> >> Think this continues from there onwards as well. I mean, even if we >> continued from the UI, at the model building step, you will notice that the >> model building is failing. >> >> The issue was with the spark filter we used to remove the header. Reason >> behind this was, spark reads each row with quotes. But the header row we >> had stored from some earlier step ignores the quotes, and was stored as >> string. So when the header removal filter compares the header row we had >> stored, against each row read by spark, return is always false. Hence the >> actual header row does not get removed at any point. This eventually breaks >> the model building as there are multiple data types in a column (due to the >> header row). >> Not sure whether this was automatically fixed after the Dataset summary >> calculation was also done using Spark, rather than Apache commons CSV >> reader. >> >> Thanks for looking in to it. >> >> >> On Mon, Jun 1, 2015 at 2:35 PM, Nirmal Fernando <[email protected]> wrote: >> >>> Ok, thanks for the clarification. >>> >>> On Mon, Jun 1, 2015 at 2:32 PM, Manorama Perera <[email protected]> >>> wrote: >>> >>>> The issue noticed is, the analysis steps cannot be proceeded after the >>>> explore step. There's no exception trace shown. >>>> >>>> The ajax request sending fails when double quotes are included in the >>>> json payload. >>>> >>>> Thanks. >>>> >>>> On Mon, Jun 1, 2015 at 2:21 PM, Nirmal Fernando <[email protected]> >>>> wrote: >>>> >>>>> Merged. Thanks. >>>>> >>>>> Can you please add the exception stack trace to the Jira too? >>>>> >>>>> On Mon, Jun 1, 2015 at 2:14 PM, Manorama Perera <[email protected]> >>>>> wrote: >>>>> >>>>>> Hi, >>>>>> >>>>>> Please find the fix [1] for JIRA [2]. >>>>>> >>>>>> [1] https://github.com/wso2/product-ml/pull/154 >>>>>> [2] https://wso2.org/jira/browse/ML-22 >>>>>> >>>>>> -- >>>>>> Manorama Perera >>>>>> Software Engineer >>>>>> WSO2, Inc.; http://wso2.com/ >>>>>> Mobile : +94716436216 >>>>>> >>>>> >>>>> >>>>> >>>>> -- >>>>> >>>>> Thanks & regards, >>>>> Nirmal >>>>> >>>>> Associate Technical Lead - Data Technologies Team, WSO2 Inc. >>>>> Mobile: +94715779733 >>>>> Blog: http://nirmalfdo.blogspot.com/ >>>>> >>>>> >>>>> >>>> >>>> >>>> -- >>>> Manorama Perera >>>> Software Engineer >>>> WSO2, Inc.; http://wso2.com/ >>>> Mobile : +94716436216 >>>> >>> >>> >>> >>> -- >>> >>> Thanks & regards, >>> Nirmal >>> >>> Associate Technical Lead - Data Technologies Team, WSO2 Inc. >>> Mobile: +94715779733 >>> Blog: http://nirmalfdo.blogspot.com/ >>> >>> >>> >> >> >> -- >> *Supun Sethunga* >> Software Engineer >> WSO2, Inc. >> http://wso2.com/ >> lean | enterprise | middleware >> Mobile : +94 716546324 >> > > > > -- > Manorama Perera > Software Engineer > WSO2, Inc.; http://wso2.com/ > Mobile : +94716436216 > -- Thanks & regards, Nirmal Associate Technical Lead - Data Technologies Team, WSO2 Inc. Mobile: +94715779733 Blog: http://nirmalfdo.blogspot.com/
_______________________________________________ Dev mailing list [email protected] http://wso2.org/cgi-bin/mailman/listinfo/dev
