TimeArray might accept Symbol with getindex ([...])

[femtotrader]

Hello,

There is a strange API difference between TimeArray (from TimeSeries.jl) and DataFrame (from DataFrames.jl)

julia> filename = "test/ford_2012.csv"
julia> ohlcv = readtimearray(filename)
250x5 TimeSeries.TimeArray{Float64,2,Date,Array{Float64,2}} 2012-01-03 to 2012-12-31

             Open     High     Low      Close    Volume
2012-01-03 | 11.0     11.25    10.99    11.13    45709900
2012-01-04 | 11.15    11.53    11.07    11.3     79725200
2012-01-05 | 11.33    11.63    11.24    11.59    67877500
2012-01-06 | 11.74    11.8     11.52    11.71    59840700
⋮
2012-12-26 | 12.31    12.79    12.31    12.79    140331900
2012-12-27 | 12.79    12.81    12.36    12.76    108315100
2012-12-28 | 12.55    12.88    12.52    12.87    95668600
2012-12-31 | 12.88    13.08    12.76    12.95    106908900

julia> ohlcv["Close"]
250x1 TimeSeries.TimeArray{Float64,1,Date,Array{Float64,1}} 2012-01-03 to 2012-12-31

             Close
2012-01-03 | 11.13
2012-01-04 | 11.3
2012-01-05 | 11.59
2012-01-06 | 11.71
⋮
2012-12-26 | 12.79
2012-12-27 | 12.76
2012-12-28 | 12.87
2012-12-31 | 12.95

with DataFrames

julia> dfOHLCV = readtable(filename);

julia> dfOHLCV[:Date] = Date(dfOHLCV[:Date]);

julia> dfOHLCV
250x6 DataFrames.DataFrame
│ Row │ Date       │ Open  │ High  │ Low   │ Close │ Volume    │
┝━━━━━┿━━━━━━━━━━━━┿━━━━━━━┿━━━━━━━┿━━━━━━━┿━━━━━━━┿━━━━━━━━━━━┥
│ 1   │ 2012-01-03 │ 11.0  │ 11.25 │ 10.99 │ 11.13 │ 45709900  │
│ 2   │ 2012-01-04 │ 11.15 │ 11.53 │ 11.07 │ 11.3  │ 79725200  │
│ 3   │ 2012-01-05 │ 11.33 │ 11.63 │ 11.24 │ 11.59 │ 67877500  │
│ 4   │ 2012-01-06 │ 11.74 │ 11.8  │ 11.52 │ 11.71 │ 59840700  │
│ 5   │ 2012-01-09 │ 11.83 │ 11.95 │ 11.7  │ 11.8  │ 53981500  │
│ 6   │ 2012-01-10 │ 12.0  │ 12.05 │ 11.63 │ 11.8  │ 121750600 │
│ 7   │ 2012-01-11 │ 11.74 │ 12.18 │ 11.65 │ 12.07 │ 63806000  │
│ 8   │ 2012-01-12 │ 12.16 │ 12.18 │ 11.89 │ 12.14 │ 48687700  │
│ 9   │ 2012-01-13 │ 12.01 │ 12.08 │ 11.84 │ 12.04 │ 46366700  │
│ 10  │ 2012-01-17 │ 12.2  │ 12.26 │ 11.96 │ 12.02 │ 44398400  │
│ 11  │ 2012-01-18 │ 12.03 │ 12.37 │ 12.0  │ 12.34 │ 47102700  │
│ 12  │ 2012-01-19 │ 12.48 │ 12.72 │ 12.43 │ 12.61 │ 70894200  │
│ 13  │ 2012-01-20 │ 12.55 │ 12.64 │ 12.45 │ 12.59 │ 43705700  │
│ 14  │ 2012-01-23 │ 12.69 │ 12.84 │ 12.55 │ 12.66 │ 49379700  │
│ 15  │ 2012-01-24 │ 12.56 │ 12.86 │ 12.46 │ 12.82 │ 45768400  │
│ 16  │ 2012-01-25 │ 12.8  │ 12.98 │ 12.7  │ 12.93 │ 54021600  │
│ 17  │ 2012-01-26 │ 13.03 │ 13.05 │ 12.66 │ 12.79 │ 75470700  │
│ 18  │ 2012-01-27 │ 11.96 │ 12.53 │ 11.79 │ 12.21 │ 142155300 │
⋮
│ 232 │ 2012-12-04 │ 11.4  │ 11.44 │ 11.23 │ 11.31 │ 37760200  │
│ 233 │ 2012-12-05 │ 11.32 │ 11.4  │ 11.18 │ 11.31 │ 33152400  │
│ 234 │ 2012-12-06 │ 11.26 │ 11.31 │ 11.19 │ 11.24 │ 31065800  │
│ 235 │ 2012-12-07 │ 11.27 │ 11.5  │ 11.26 │ 11.48 │ 38404500  │
│ 236 │ 2012-12-10 │ 11.41 │ 11.53 │ 11.41 │ 11.47 │ 26025200  │
│ 237 │ 2012-12-11 │ 11.51 │ 11.58 │ 11.4  │ 11.49 │ 36326900  │
│ 238 │ 2012-12-12 │ 11.52 │ 11.56 │ 11.43 │ 11.47 │ 31099900  │
│ 239 │ 2012-12-13 │ 11.46 │ 11.5  │ 11.21 │ 11.27 │ 35443200  │
│ 240 │ 2012-12-14 │ 11.27 │ 11.27 │ 11.03 │ 11.1  │ 36933500  │
│ 241 │ 2012-12-17 │ 11.16 │ 11.41 │ 11.14 │ 11.39 │ 46983300  │
│ 242 │ 2012-12-18 │ 11.48 │ 11.68 │ 11.4  │ 11.67 │ 61810400  │
│ 243 │ 2012-12-19 │ 11.79 │ 11.85 │ 11.62 │ 11.73 │ 54884700  │
│ 244 │ 2012-12-20 │ 11.74 │ 11.8  │ 11.58 │ 11.77 │ 47750100  │
│ 245 │ 2012-12-21 │ 11.55 │ 11.86 │ 11.47 │ 11.86 │ 94489300  │
│ 246 │ 2012-12-24 │ 11.67 │ 12.4  │ 11.67 │ 12.4  │ 91734900  │
│ 247 │ 2012-12-26 │ 12.31 │ 12.79 │ 12.31 │ 12.79 │ 140331900 │
│ 248 │ 2012-12-27 │ 12.79 │ 12.81 │ 12.36 │ 12.76 │ 108315100 │
│ 249 │ 2012-12-28 │ 12.55 │ 12.88 │ 12.52 │ 12.87 │ 95668600  │
│ 250 │ 2012-12-31 │ 12.88 │ 13.08 │ 12.76 │ 12.95 │ 106908900 │

julia> dfOHLCV[:Open]
250-element DataArrays.DataArray{Float64,1}:
 11.0
 11.15
 11.33
 11.74
 11.83
 12.0
 11.74
 12.16
 12.01
 12.2
 12.03
 12.48
 12.55
 12.69
 12.56
 12.8
 13.03
 11.96
 12.06
  ⋮
 11.4
 11.32
 11.26
 11.27
 11.41
 11.51
 11.52
 11.46
 11.27
 11.16
 11.48
 11.79
 11.74
 11.55
 11.67
 12.31
 12.79
 12.55
 12.88

but

julia> dfOHLCV["Close"]
ERROR: MethodError: `getindex` has no method matching getindex(::DataFrames.DataFrame, ::ASCIIString)
Closest candidates are:
  getindex(::DataFrames.DataFrame, ::Real, ::Union{Real,Symbol})
  getindex{T<:Union{Real,Symbol}}(::DataFrames.DataFrame, ::Real, ::AbstractArray{T<:Union{Real,Symbol},1})
  getindex(::DataFrames.DataFrame, ::Real, ::Colon)
  ...

julia> ohlcv[:Close]
ERROR: MethodError: `getindex` has no method matching getindex(::TimeSeries.TimeArray{Float64,2,Date,Array{Float64,2}}, ::Symbol)
Closest candidates are:
  getindex{T,N,D}(::TimeSeries.TimeArray{T,N,D,A<:AbstractArray{T,N}}, ::Int64)
  getindex{T,N,D}(::TimeSeries.TimeArray{T,N,D,A<:AbstractArray{T,N}}, ::UnitRange{Int64})
  getindex{T,N,D}(::TimeSeries.TimeArray{T,N,D,A<:AbstractArray{T,N}}, ::Array{Int64,1})
  ...

It will be nice if getindex could accept Symbol as second parameter when a TimeArray is given as first parameter.

Kind regards

PS : see JuliaData/DataFrames.jl#959

[nalimilan]

Data frames use symbols as column names in order to allow accessing them using df.col when overloading this syntax becomes possible (JuliaLang/julia#1974). So I would encourage TimeSeries to do the same.

[milktrader]

Yep, I'm on board with this.

[milktrader]

I'm also open to changing TimeArray to TimeFrame, but this requires a bit more discussion. It would be natural (I think) to make this important change at the same time as changing the data structure's name to TimeFrame.

[iblislin]

I want to fully switch to Symbol, and then introduce the xts's filtering mentioned by @dysonance

[iblislin]

getproperty/setproperty! merged into master:
JuliaLang/julia#24960