Consider adding a parameter to nest aggregated columns: aggregate(under: String?)

[koperagen]

df.groupBy { key }.aggregate {
  1 into "col"
}

schema:

key
col

df.groupBy { key }.aggregate(under = "statistics") {
  1 into "col"
}

schema:

key
statistics:
  col

Make sure it's consistent across other aggregate overloads (if makes sense)

[koperagen]

This creates a ColumnGroup by default. It's a bit different from aggregate with into statements

df_3.groupBy { year }.aggregate { 
    maxBy { gflops.toDouble() }
}

year
aggregated:
  other columns

[zaleslaw]

Unclear description of feature, need to refresh together @koperagen