Edit on GitHub

Aggregation

Sampling

Weekdays

Temporal.mondays — Function
mondays(x::TS)

Return a time series containing all observations occuring on Mondays of the given TS object.

source
Temporal.tuesdays — Function
tuesdays(x::TS)

Return a time series containing all observations occuring on Tuesdays of the given TS object.

source
Temporal.wednesdays — Function
wednesdays(x::TS)

Return a time series containing all observations occuring on Wednesdays of the given TS object.

source
Temporal.thursdays — Function
thursdays(x::TS)

Return a time series containing all observations occuring on Thursdays of the given TS object.

source
Temporal.fridays — Function
fridays(x::TS)

Return a time series containing all observations occuring on Fridays of the given TS object.

source
Temporal.saturdays — Function
saturdays(x::TS)

Return a time series containing all observations occuring on Saturdays of the given TS object.

source
Temporal.sundays — Function
sundays(x::TS)

Return a time series containing all observations occuring on Sundays of the given TS object.

source

Interval Boundaries

Weekly

Temporal.bow — Function
bow(x::TS; cal::Bool=false)

Return a time series containing all observations occuring at the beginnings of the weeks of the input.

  • If cal is false, only observations occurring the last calendar day of the week are returned
  • If cal is true, all observation for which the previous index is a prior week are returned
source
Temporal.eow — Function
eow(x::TS; cal::Bool=false)

Return a time series containing all observations occuring at the ends of the weeks of the input.

  • If cal is false, only observations occurring the last calendar day of the week are returned
  • If cal is true, all observation for which the next index is a new week are returned
source

Monthly

Temporal.bom — Function
bom(x::TS; cal::Bool=false)

Return a time series containing all observations occuring at the beginnings of the months of the input.

  • If cal is false, only observations occurring the last calendar day of the month are returned
  • If cal is true, all observation for which the previous index is a prior month are returned
source
Temporal.eom — Function
eom(x::TS; cal::Bool=false)

Return a time series containing all observations occuring at the ends of the months of the input.

  • If cal is false, only observations occurring the last calendar day of the month are returned
  • If cal is true, all observation for which the next index is a new month are returned
source

Quarterly

Temporal.boq — Function
boq(x::TS; cal::Bool=false)

Return a time series containing all observations occuring at the beginnings of the quarters of the input.

  • If cal is false, only observations occurring the last calendar day of the quarter are returned
  • If cal is true, all observation for which the previous index is a prior quarter are returned
source
Temporal.eoq — Function
eoq(x::TS; cal::Bool=false)

Return a time series containing all observations occuring at the ends of the quarters of the input.

  • If cal is false, only observations occurring the last calendar day of the quarter are returned
  • If cal is true, all observation for which the next index is a new quarter are returned
source

Yearly

Temporal.boy — Function
boy(x::TS; cal::Bool=false)

Return a time series containing all observations occuring at the beginnings of the years of the input.

  • If cal is false, only observations occurring the last calendar day of the year are returned
  • If cal is true, all observation for which the previous index is a prior year are returned
source
Temporal.eoy — Function
eoy(x::TS; cal::Bool=false)

Return a time series containing all observations occuring at the ends of the years of the input.

  • If cal is false, only observations occurring the last calendar day of the year are returned
  • If cal is true, all observation for which the next index is a new year are returned
source

Collapsing

Temporal.collapse — Function
collapse(x::TS, at::Function; fun::Function=last, args...)::TS

Compute the function fun over period boundaries defined by at (e.g. eom, boq, etc.) and returng a time series of the results.

Keyword arguments (args...) are passed to the function fun.

source
julia> using Temporal, Statistics, Dates

julia> X = TS(randn(100, 4))
100x4 TS{Float64,Dates.Date}: 2019-10-15 to 2020-01-22

Index       A        B        C        D
2019-10-15  -0.1866  -0.84    -0.1397  2.2043
2019-10-16  0.7517   -0.8447  -0.5989  -1.4443
2019-10-17  -1.8092  0.0303   0.6289   -1.8995
2019-10-18  -1.1157  1.3227   1.0881   0.283
2019-10-19  2.3262   -1.0653  0.8624   -1.4475
2019-10-20  -1.3373  -0.8158  -0.7763  1.7431
2019-10-21  -1.4505  0.7358   -1.1379  -0.2587
⋮
2020-01-15  -1.2701  0.0881   0.3331   0.3654
2020-01-16  -0.547   -0.656   0.3009   1.12
2020-01-17  -0.0798  -1.5292  -1.2675  0.8403
2020-01-18  1.0677   -0.1692  -0.229   -0.8929
2020-01-19  0.2681   -0.5048  0.569    -0.0101
2020-01-20  2.202    0.171    0.1568   -0.8425
2020-01-21  1.1247   -0.882   -1.7938  1.926
2020-01-22  0.0681   -1.7146  1.472    -0.6442

julia> collapse(X, eom, fun=mean)
3x4 TS{Float64,Dates.Date}: 2019-10-31 to 2019-12-31

Index       A        B        C        D
2019-10-31  0.1631   0.0517   0.0729   -0.1831
2019-11-30  -0.0219  0.0558   0.4351   0.0572
2019-12-31  -0.0135  -0.0714  -0.0587  0.1542

julia> last_month = string(X.index[end])[1:7]
"2020-01"

julia> mean(X[last_month])
-0.0264185897272492

Summarizing

Temporal.apply — Function
apply(x::TS, dim::Int=1; fun::Function=sum)

Apply function fun across dimension dim of a time series x.

  • If dim is 1, then apply fun to each row of x, returning a time series (TS) object of the results
  • If dim is 2, then apply fun to each column of x, returning an Array of the results
source
julia> using Temporal, Statistics

julia> X = TS(randn(100, 4))
100x4 TS{Float64,Dates.Date}: 2019-10-15 to 2020-01-22

Index       A        B        C        D
2019-10-15  -0.9595  0.1335   -1.0634  0.2979
2019-10-16  -0.5834  1.2348   -0.744   0.2918
2019-10-17  1.2931   -1.4254  -0.653   -0.2039
2019-10-18  0.8427   1.3852   0.8431   0.243
2019-10-19  -0.8869  1.131    -0.2246  0.6173
2019-10-20  1.2795   0.7889   -1.2842  1.0755
2019-10-21  -2.2764  1.0616   0.6205   -0.0099
⋮
2020-01-15  -0.5043  -1.3508  -0.1313  -0.9573
2020-01-16  -1.485   1.021    -1.9404  0.9268
2020-01-17  0.8451   -1.4235  0.3153   -0.0515
2020-01-18  -0.1652  0.1091   -0.1821  -2.225
2020-01-19  -0.2702  1.1492   0.7934   1.5887
2020-01-20  -0.2894  0.3906   0.817    0.9331
2020-01-21  -0.4678  -0.7675  -1.7152  0.4977
2020-01-22  -0.2223  -0.5855  -0.5913  -0.5921

julia> apply(X, 1, fun=sum)
100x1 TS{Float64,Dates.Date}: 2019-10-15 to 2020-01-22

Index       Sum
2019-10-15  -1.5915
2019-10-16  0.1992
2019-10-17  -0.9892
2019-10-18  3.3139
2019-10-19  0.6368
2019-10-20  1.8597
2019-10-21  -0.6042
⋮
2020-01-15  -2.9437
2020-01-16  -1.4776
2020-01-17  -0.3146
2020-01-18  -2.4633
2020-01-19  3.261
2020-01-20  1.8512
2020-01-21  -2.4528
2020-01-22  -1.9912

julia> apply(X, 2, fun=sum)
1×4 LinearAlgebra.Adjoint{Float64,Array{Float64,1}}:
 -8.12431  5.90707  9.15881  0.403375