Skip to content
This repository has been archived by the owner on Jan 28, 2023. It is now read-only.

Should have FloatCol? #101

Open
wangtieqiao opened this issue Nov 3, 2020 · 2 comments
Open

Should have FloatCol? #101

wangtieqiao opened this issue Nov 3, 2020 · 2 comments

Comments

@wangtieqiao
Copy link

In the case to use Krangl for data pre-processing in Deep Learning (model operate on Float), to have FloatCol might be useful for large data set (memory usage). -- only see DoubleCol implemented.

@holgerbrandl
Copy link
Owner

I share the argument. Technically it would be even easy to do so. However. as we are currently use Array<Double?> to support NA, DoubleCol is not backed by a double-array, so we loose a lot of efficiency. The same would apply for FloatCol, so I'm not sure if adding using the current backend storage model makes much sense wrt to memory usage.

So my plan was/is to first find a more memory-efficient DoubleCol implementation, and then adopt the pattern when adding FloatCol.

Any help is welcome here.

@holgerbrandl
Copy link
Owner

See

class DoubleCol(name: String, val values: Array<Double?>) : DataCol(name) {

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Projects
None yet
Development

No branches or pull requests

2 participants