filter, map, but no reduce, no group

[alexec]

It would be good to complete this set of operations.

I think group would be:

group(Prices,  {# > 10})

But I'm not sure the closure syntax can support reduce op due to the need for the closure to take two args, something like

reduce(Prices, 0, {.value + .acc})

[antonmedv]

Yes, I’m also thinking about reduce built in for some time. Two arguments maybe done with for example #1, #2 or #acc named parameters.

But what should group func do?

[alexec]

Group should partition an array into other arrays based on a key. Like SQL group by

[antonmedv]

So it should return a map?

[alexec]

I'm not sure. It could be an array of arrays, but the downside would be that you don't then know what the grouping key was.

I think the ultimate litmus test would be - can you do Hadoop style map-reduce?

https://techmytalk.com/2014/11/07/hadoop-mapreduce-group-by-operation-part1/#:~:text=The%20grouping%20of%20map%20output,the%20Map%20to%20the%20Reducer.

[alexec]

@antonmedv I'm not sure I'm skilled enough to write these on my own, could we get one a Zoom and you show me how please?

[mhr3]

+1 for reduce, perhaps more than one param could be done by extending closures with a js-like "header" (x,y,z) => {...}, where the header is optional, and when not present basically becomes (#) => ?

[antonmedv]

Yes, also was thinking about (params) => {...} style. But what about {x, y => ...}?

[mhr3]

Looks like a closure returning two values: x and a bool :)

[antonmedv]

{#1 >= #2}

[pritamdey12]

@antonmedv Hi, any update on the reduce function?

[antonmedv]

Nope.

[forrestlinfeng]

So it should return a map?

yes, this feature is useful like as slice to map。

[antonmedv]

Added groupBy func.

groupBy(1..9, # % 2)[1]

[antonmedv]

Added reduce

reduce(1..9, # + #acc)
reduce(1..9, # + #acc, 0)