inline functions just let you inline the real type-param being used in place of the generic one

@hoang Erasure is a much better alternative than reification: http://beust.com/erasure 2 replies Define "works well"...

I love the idioms page in the documentation

I do not, however like the map literal syntax, it makes no sense to say <key> to <values> ? shouldn’t it be values to key?

why doesn't that make sense?

key => value

is an idiom that exists pretty much everywhere

my notion of a map is that the key is a container for a set of values

the english word to implies something being associate to its referent, not the referent being affixed to the collection

that’s all I am saying, might be better to use by, then it would be really clear that you were saying ‘I wish to access this collection by this key’

Dunno makes sense to me: "x [maps] to y"

I've always used "to" in my head before kotlin, but perhaps I'm weird like that

yeah I am not saying the world should conform to my way of thinking, was wondering if anyone else thought it was counterintuitive..

the key points to a value. In other languages, map literals would be

key => value

,

key: value

,

key=value

, etc. I think

key to value

just feels right here,

value to key

is not very intuitive to me

yeah ok

value at key

and another two-letter function! 🙂

I prefer that 🙂 and 2 letter functions are fine by me

@codeslubber you know how to create such an infix function, don’t you? 🙂

haha I will do it good exercise..

funny found this immediately: https://discuss.kotlinlang.org/t/infix-functions/1076

@thecipster has left the channel

Started to play with Pandas today. It would be nice to see a DataFrame implementation for Kotlin get momentum. https://github.com/holgerbrandl/krangl

@thomasnield @khud you guys need to talk 🙂

@khud I think @holgerbrandl would be interested as well

@orangy @holgerbrandl @thomasnield I guess Kotlin needs ML stuff (dataframes, models and so on) but may be in some uniform way. Moreover, we should think about desktop ML/Spark (cluster) integration issues. Meanwhile I am working on Kotlin/Spark integration.

That’s what I suggest to talk about. You all seem to be interested in data stuff, you may cooperate and build a nice data science community around/with Kotlin. I see @holgerbrandl even did an R plugin for IJ, so there is even more hope. Of course, we at JetBrains will help as much as we can.

@orangy 👍

I did use Spark 1.6 with Kotlin, no issues, just a couple of rough patches

If you need standalone production job, you can use Java Spark API almost without any problems. But, I think some wrappers are better choice. If you want to do some ad hoc computations you need to have a REPL.

I end writing several extension functions to have a more idiomatic experience, but yes, REPL will be a welcomed option

I have a nice Spark Shell for Kotlin and will share it asap. I also wrote many wrappers with good Kotlin signatures but need some time to finish.

Is there any good open source API project written in spark-java(java/kotlin) with good architecture(like android-cleanarchitecture, MVP/MVVM examples). I'd see only random example posts code scattered through out Github? (I'm kinda new to developing API's, I have no idea how to organize code)!