Github user mengxr commented on the pull request:
https://github.com/apache/spark/pull/9595#issuecomment-173378701
LIBSVM is a text format and hence we need to consider the cost of storing
numerical values. In the current implementation, the output could be some text
like `1 1:0.12345678901234`, where people might not need more than 6 digits on
the feature values. `1 1:0.123456` should be sufficient. But we don't have an
option to control the formatting of values in the implementation. It could be
just a parameter of this LIBSVM data source, or it could be a global flag
because the same issue also applies to other text formats like CSV and JSON.
Let's create a JIRA and move our discussion there.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
