Currently, the CSV parser supports the following six data types: bool, integer, string, float, double, and date. This parser does not support data type inference, so you should give the right setting for each column in the input schema definition.
For the input CSV file, adjacent fields must be separated by a single comma (other delimiters are not supported), and any field can be quoted.
Note
The nested comma must be quoted, and the embedded double-quote characters and line breaks is not being supported so far. Also, the leading and trailing spaces between fields, as well as the heading record is forbidden.
For output data, the CSV parser will allow you to select some columns, and only the selected columns would be outputted by the object-stream interface.