feat: make .topk() and .value_counts() output same column names

[NickCrews]

Is your feature request related to a problem?

Ideally, these two methods would return the same data, except that .topk() is .order_by()ed and .limit()ed.

But currently the output column names are different (a_count vs Count(a))

import ibis

ibis.options.interactive = True

t = ibis.memtable({"a": [1, 1, 2, None]})
t.a.value_counts()
# ┏━━━━━━━━━┳━━━━━━━━━┓
# ┃ a       ┃ a_count ┃
# ┡━━━━━━━━━╇━━━━━━━━━┩
# │ float64 │ int64   │
# ├─────────┼─────────┤
# │     2.0 │       1 │
# │    NULL │       1 │
# │     1.0 │       2 │
# └─────────┴─────────┘
t.a.topk(10)
# ┏━━━━━━━━━┳━━━━━━━━━━┓
# ┃ a       ┃ Count(a) ┃
# ┡━━━━━━━━━╇━━━━━━━━━━┩
# │ float64 │ int64    │
# ├─────────┼──────────┤
# │     1.0 │        2 │
# │     2.0 │        1 │
# │    NULL │        0 │
# └─────────┴──────────┘

Describe the solution you'd like

I think adding a param name, with the same semantics as lname and rname in join, with a default of {name}_count, would be nice. Or, just making this always be the behavior, with no option, would also be ok. I vote for {name}_count over Count({name}), because it is easier to construct the first one programatically, eg

cols = ["a", "b", "c"]
counts = [t[col].value_counts().filter(_[col + "_count"] > 0)]

Also this is more consistent with our default behavior of adding suffixes in .join()

Anything else we could do to make these two methods have the same API would be great, but I think it doesn't make sense to unify them any more.

What version of ibis are you running?

main

What backend(s) are you using, if any?

No response

Code of Conduct

• I agree to follow this project's Code of Conduct