The description of that property from the docs says... "Specifies a String that, if present as a value in the CSV, should be considered a null field instead of using the literal value."
I believe the use-case is if you had a CSV line like "a, b, NULL, d" which is different from "a, b, , d". If you want to treat the first one with the string "NULL" as an actual null value, then you can specify "NULL" in the "Null String" property to tell the reader. On Tue, Dec 19, 2017 at 2:40 PM, Aruna Sankaralingam < [email protected]> wrote: > Thank you Bryan. I updated the schema and that worked. But could you > please let me know the purpose of this property in CSVReader – “NULL > String”? > > > > *From:* Bryan Bende [mailto:[email protected]] > *Sent:* Tuesday, December 19, 2017 1:57 PM > *To:* [email protected] > *Subject:* Re: NULL String > > > > Hello, > > What does your schema look like? > > You would need the schema to indicate that the field is nullable like this: > > "type": ["string", "null"] } > > If you only have "type": ["string"] } then it will produce an error when > reading a null value. > > -Bryan > > > > On Tue, Dec 19, 2017 at 1:46 PM, Aruna Sankaralingam < > [email protected]> wrote: > > I have a CSV file in which some of the fields have NULL values. I am > getting this error that “Field cannot be null”. How do I let Nifi know to > accept NULL values? > > > > I see this property in the CSV reader called “NULL String”. Is there > anything that I can give here? > > > > >
