Hi, Generally speaking, I think it is preferable to make system more tolerant to input data so that it is simpler to use from user point of view. And I think it is safe to align with other format like ORC and parquet.
Regards, Jacky > 在 2017年5月17日,下午12:30,manish gupta <tomanishgupt...@gmail.com> 写道: > > Hi Vin, > > Value is getting displayed as null because data is not getting trimmed. By > default carbondata will not modify any data. But there is option to verify > your data. > > 1. While loading the data set the property " > BAD_RECORDS_LOGGER_ENABLE=true". This will create a bad record logger where > in you can see what all records are invalid. > > 2. Secondly by default trimming data is disabled in the code. > > parserSettings.setIgnoreLeadingWhitespaces(false); > parserSettings.setIgnoreTrailingWhitespaces(false); > > > We can provide an option during data load to trim whitespaces and keep the > default value as false to keep the current behavior intact. Example > load data inpath 'D:/my_user2.csv' into table my_user_c > OPTIONS('TRIM_WHITESPACE'='TRUE') > > @Dev team please share your inputs for the 2nd option. > > Regards > Manish Gupta > > > On Wed, May 17, 2017 at 7:47 AM, David Cai <david.c...@gmail.com> wrote: > >> I suggest to trim white space before converting a string value to a numeric >> value. >> >> >> >> ----- >> Best Regards >> David Cai >> -- >> View this message in context: http://apache-carbondata-dev- >> mailing-list-archive.1130556.n5.nabble.com/White-Space- >> does-not-trim-in-CarbonData-tp12736p12751.html >> Sent from the Apache CarbonData Dev Mailing List archive mailing list >> archive at Nabble.com. >>