Parquet to TimeSeries TimeArray

Hello,

I'm reading a Parquet file which was previously saved with Python Pandas DataFrame and try to open it with Julia as TimeSeries TimeArray.

I noticed

ds = Parquet2.Dataset("/path/to/file")
dfc = ds |> TableOperations.select(:time, :open, :high, :low, :close, :volume) |> DataFrame

can help to convert to DataFrame but

using TimeSeries
ds |> TableOperations.select(:time, :open, :high, :low, :close, :volume) |> TimeArray

raises

ERROR: UndefKeywordError: keyword argument `timestamp` not assigned
Stacktrace:
 [1] TimeArray(x::TableOperations.Select{Tables.CopiedColumns{NamedTuple{(:time, :open, :high, :low, :close, :volume, :close_time, :quote_asset_volume, :number_of_trades, :taker_buy_base_asset_volume, :taker_buy_quote_asset_volume, :ignore), Tuple{Vector{Union{Missing, DateTime}}, Vector{Union{Missing, Float64}}, Vector{Union{Missing, Float64}}, Vector{Union{Missing, Float64}}, Vector{Union{Missing, Float64}}, Vector{Union{Missing, Float64}}, Vector{Union{Missing, DateTime}}, Vector{Union{Missing, Float64}}, Vector{Union{Missing, Int64}}, Vector{Union{Missing, Float64}}, Vector{Union{Missing, Float64}}, Vector{Union{Missing, Float64}}}}}, true, (:time, :open, :high, :low, :close, :volume)})
   @ TimeSeries.TablesIntegration C:\Users\scell\.julia\packages\TimeSeries\KZKvJ\src\tables.jl:71
 [2] |>(x::TableOperations.Select{Tables.CopiedColumns{NamedTuple{(:time, :open, :high, :low, :close, :volume, :close_time, :quote_asset_volume, :number_of_trades, :taker_buy_base_asset_volume, :taker_buy_quote_asset_volume, :ignore), Tuple{Vector{Union{Missing, DateTime}}, Vector{Union{Missing, Float64}}, Vector{Union{Missing, Float64}}, Vector{Union{Missing, Float64}}, Vector{Union{Missing, Float64}}, Vector{Union{Missing, Float64}}, Vector{Union{Missing, DateTime}}, Vector{Union{Missing, Float64}}, Vector{Union{Missing, Int64}}, Vector{Union{Missing, Float64}}, Vector{Union{Missing, Float64}}, Vector{Union{Missing, Float64}}}}}, true, (:time, :open, :high, :low, :close, :volume)}, f::Type{TimeArray})
   @ Base .\operators.jl:907
 [3] top-level scope
   @ REPL[108]:1

What is the most appropriate way of converting to TimeSeries ?

TimeArray(ds |> TableOperations.select(:time, :open, :high, :low, :close, :volume), timestamp=:time)

but is odd as it mixes function chaining operator (|>) and function call.

So maybe something like

ds |> TableOperations.select(:time, :open, :high, :low, :close, :volume) |> t -> TimeArray(t, timestamp=:time)

looks better

Isn't there a way for select to behave as SQL alias (SELECT column1 AS c1, column2 FROM table) ?

A nicer API call could be

ds |> TableOperations.select(:time => :timestamp, :open, :high, :low, :close, :volume) |> TimeArray

Any opinion? https://github.com/JuliaData/TableOperations.jl/issues/33

Kind regards