There are much more fearsome Dataframe wizards than I am on this forum, but
I would use something like this:
using RDatasets
df = dataset("datasets", "iris")
by(df, :PetalWidth) do subdf
subdf[indmin(subdf[:SepalLength]), :]
end
although that's keeping the entries with the lowest sum, and discarding
everything else. Could you clarify what you mean by "I try to find a simple
way to remove the duplicates that have the lowest sum without changing the
values of b and c."?
On Friday, May 27, 2016 at 1:56:54 PM UTC-4, Fred wrote:
>
> Hi,
>
> I have a dataframe df2 and the last column is the sum = b+c :
>
> julia> df2
> 8x4 DataFrames.DataFrame
> │ Row │ a │ b │ c │ sum │
> ┝━━━━━┿━━━┿━━━┿━━━━━━━━━━━┿━━━━━━━━━━━┥
> │ 1 │ 1 │ 2 │ -0.163564 │ 1.83644 │
> │ 2 │ 2 │ 1 │ 0.731038 │ 1.73104 │
> │ 3 │ 3 │ 2 │ 0.0951149 │ 2.09511 │
> │ 4 │ 4 │ 1 │ 0.195321 │ 1.19532 │
> │ 5 │ 1 │ 2 │ 1.97058 │ 3.97058 │
> │ 6 │ 2 │ 1 │ 0.150826 │ 1.15083 │
> │ 7 │ 3 │ 2 │ 0.422046 │ 2.42205 │
> │ 8 │ 4 │ 1 │ -1.36549 │ -0.365486 │
>
> we can se that the column a have duplicates (1,2,3). I try to find a
> simple way to remove the duplicates that have the lowest sum without
> changing the values of b and c.
>
> I tried :
>
> julia> aggregate(df2, :a, maximum)
>
>
> 4x4 DataFrames.DataFrame
> │ Row │ a │ b_maximum │ c_maximum │ sum_maximum │
> ┝━━━━━┿━━━┿━━━━━━━━━━━┿━━━━━━━━━━━┿━━━━━━━━━━━━━┥
> │ 1 │ 1 │ 2 │ 1.97058 │ 3.97058 │
> │ 2 │ 2 │ 1 │ 0.731038 │ 1.73104 │
> │ 3 │ 3 │ 2 │ 0.422046 │ 2.42205 │
> │ 4 │ 4 │ 1 │ 0.195321 │ 1.19532 │
>
>
>
> but this is wrong because I don't want b_maximum and c_maximum but only
> sum_maximum :
>
>
> Row │ a │ b │ c │ sum_maximum
>
>
> I don't think that there is a simple way to do that but I ask the question
> in case ;)
>
> Thank you very much in advance !
>
