Function 'explode' should work with other collections like Set, Arrays, etc. not only Lists

[belovrv]

[Jolanrensen]

So, all iterables? (+ arrays)

[belovrv]

Yep, I think so

[zaleslaw]

I will take this one, it will help me to understand something in the framework and its functions

[zaleslaw]

Hi, @Jolanrensen @belovrv the task description is little bit unclear for me.

What does mean to support work with Set or Arrays in that context?

For example we have a test for explode function:

    @Test
    fun `explode into`() {
        val df = dataFrameOf("a" to listOf(1), "b" to listOf(listOf(2, 3)))
        val exploded = df.explode { "b" into "c" }
        val expected = dataFrameOf("a" to listOf(1, 1), "c" to listOf(2, 3))
        exploded shouldBe expected
    }

Does it mean that for now we want to support the following functionality?

    @Test
    fun `explode into set`() {
        val df = dataFrameOf("a" to setOf(1), "b" to setOf(setOf(2, 3)))
        val exploded = df.explode { "b" into "c" }
        val expected = dataFrameOf("a" to setOf(1, 1), "c" to setOf(2, 3))
        exploded shouldBe expected
    }

But should we keep a type for the result or should we convert all the stuff to one type, List for example?

[Jolanrensen]

I can explain the issue with two screenshots. Let's first create two DataFrames, one with a column containing two values in a list and the second dataframe having a column containing two values inside a set:

You'd expect exploding would work for both cases since they are displayed the same but no.

[zaleslaw]

@Jolanrensen you mean low-level replacement of list on set, but what about high-level composition?

setOf(listOf) or seOf(setOf) or arrayOf(setOf())?

[Jolanrensen]

@zaleslaw You mean replacing the dataframeOf function to also allow pairs of string to sets and string to arrays? That's outside the scope of the issue.
(remember dataframeOf(vararg Pair<String, List<Any?>>) is just a function to create a dataframe with multiple columns, it has nothing directly to do with the explode function)

Otherwise, If you mean sets of lists inside a column, then it should behave the same as lists of lists in columns would behave. Which in our case means just unpacking the top-layer:


Again, in the bottom picture, the bottom result should be the same as the first one.