feat: JSON-to-CSV - improve list parsing to better support datacubes

[nathan-stender]

Currently, when two elements are not in the same root, we always cross the parsed datasets.

ASM datacubes store the x/y dimensions in lists in separate paths, which means they get crossed, so that for:

data : {
  dimensions: [[1,2]],
  measures: [[0.25, 0.5] 
}

you get

1,0.25
1,0.5
2,0.25
2,0.5

Now, we detect if all parsed datasets being combined have the same length, and if so, just concat them.

This gets us the expected result:

1,0.25
2,0.5

for datacube extraction, while still crossing with downstream nested values