JuliaTS

Implements some time series and relational algebra operations over NDSparseData.

construct

TArray(colpairs::Pair...)
TArray(keynames::Tuple, colpairs::Pair...)
TArray(data::NDSparse, keynames::Tuple, valnames::Tuple)
TArray(from::TArray, colpairs::Pair...)
TArray(from::TimeSeries.TimeArray)
TArray(keynames::Tuple, from::TimeSeries.TimeArray)
TArray(keynames::Tuple, from::DataFrames.DataFrame)

get/set

Use Julia getindex and setindex! with single or range of values to match with every dimension. E.g.:

val = ta[DateTime("2016-01-01"):DateTime("2016-01-02"), [1, 2, 3], :open]
ta[DateTime("2016-01-01"), 1, :open] = (11.2, 31.2, 21.2, 30.0)

Get column vector: getvals(ta::TArray, colname) Set column vector: setvals(ta::TArray, colname, vals::Vector) Any column can be fetched by get, but set can be used only on value columns (keys can not be set).

set operations

project(ta::TArray, keynames, valnames)
union(ta1::TArray, ta2::TArray)
intersect(ta1::TArray, ta2::TArray)
difference(ta1::TArray, ta2::TArray)

rename columns

rename(ta::TArray, keynames, valnames)
rename!(ta::TArray, keynames, valnames)
rename(ta::TArray, pfx)
rename!(ta::TArray, pfx)

groupby

# fn: the aggregator function, called per group-column
# fns: tuple of functions, one per value column, when different aggregation functions are neeed for each column
groupby(ta::TArray, keynames, fn)
groupby(ta::TArray, keynames, fns::Tuple{Function})

join

# jointype can be :inner, :left, :right
# only inner join supported now
naturaljoin(taout::TArray, ta1::TArray, ta2::TArray, jointype, fn::Function)

shift

timeshift(ta::TArray, col, by)
timeshift!(ta::TArray, col, by)

window aggregation

# fn: the aggregaion function
window(taout::TArray, ta::TArray, wspec::Window, wintype, fn)

# window specification:
immutable Window{K,S,W}
    first::K        # start of first window
    last::K         # end of the last window
    step::S         # increment window start
    width::W        # window width
end