Hi, Matt Sun from the Hadoop team has proposed a change to the way we parse tokens (CSV-196 [1]). Hadoop needs to know the raw character data which has been read from the input stream. So they propose to add an additional buffer to the Token class to hold that data.
Please comment in the ticket what you think about this change before we go in that direction, since I’m very cautions about touching the parsing code. Thank you! Benedikt [1]: https://issues.apache.org/jira/browse/CSV-196 --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
