Align a data.table onto a nanotime vector

Description

align returns the subset of data.table x that aligns on the temporal vector y

Usage

align(x, y, start, end, ...)

## S4 method for signature 'data.table,nanotime,nanoduration,nanoduration'
align(
  x,
  y,
  start = as.nanoduration(0),
  end = as.nanoduration(0),
  sopen = FALSE,
  eopen = TRUE,
  func = NULL
)

## S4 method for signature 'data.table,nanotime,missing,missing'
align(
  x,
  y,
  start = as.nanoduration(0),
  end = as.nanoduration(0),
  sopen = FALSE,
  eopen = TRUE,
  func = NULL
)

## S4 method for signature 'data.table,nanotime,nanoduration,missing'
align(
  x,
  y,
  start = as.nanoduration(0),
  end = as.nanoduration(0),
  sopen = FALSE,
  eopen = TRUE,
  func = NULL
)

## S4 method for signature 'data.table,nanotime,missing,nanoduration'
align(
  x,
  y,
  start = as.nanoduration(0),
  end = as.nanoduration(0),
  sopen = FALSE,
  eopen = TRUE,
  func = NULL
)

## S4 method for signature 'data.table,nanotime,nanoperiod,nanoperiod'
align(
  x,
  y,
  start = as.nanoperiod(0),
  end = as.nanoperiod(0),
  sopen = FALSE,
  eopen = TRUE,
  tz,
  func = NULL
)

## S4 method for signature 'data.table,nanotime,nanoperiod,missing'
align(
  x,
  y,
  start = as.nanoperiod(0),
  end = as.nanoperiod(0),
  sopen = FALSE,
  eopen = TRUE,
  tz,
  func = NULL
)

## S4 method for signature 'data.table,nanotime,missing,nanoperiod'
align(
  x,
  y,
  start = as.nanoperiod(0),
  end = as.nanoperiod(0),
  sopen = FALSE,
  eopen = TRUE,
  tz,
  func = NULL
)

Arguments

Details

For each element in y, intervals are created around this element with start and end. All the elements of x that fall within this interval are given as argument to the function func. The function func show reduce this data.frame to one unique row that will be associated with the nanotime value in y.

Value

a data.table time-series of the same length as y; this is a subset of x with the nanotime index of y

Examples

## Not run: 
y <- nanotime((1:10)*1e9)
x <- data.table(index=nanotime((1:10)*1e9), data=1:10)
align(x, y, as.nanoduration(-1e9), as.nanoduration(1e9), colMeans)

## End(Not run)

Get the index of the alignment of one vector onto another

Description

align_idx returns the index of the alignment of x on y

Usage

align_idx(x, y, start, end, ...)

## S4 method for signature 'nanotime,nanotime,nanoduration,nanoduration'
align_idx(
  x,
  y,
  start,
  end,
  sopen = FALSE,
  eopen = TRUE,
  bypass_x_check = FALSE,
  bypass_y_check = FALSE
)

## S4 method for signature 'nanotime,nanotime,missing,missing'
align_idx(
  x,
  y,
  start,
  end,
  sopen = FALSE,
  eopen = TRUE,
  bypass_x_check = FALSE,
  bypass_y_check = FALSE
)

## S4 method for signature 'nanotime,nanotime,missing,nanoduration'
align_idx(
  x,
  y,
  start,
  end,
  sopen = FALSE,
  eopen = TRUE,
  bypass_x_check = FALSE,
  bypass_y_check = FALSE
)

## S4 method for signature 'nanotime,nanotime,nanoduration,missing'
align_idx(
  x,
  y,
  start,
  end,
  sopen = FALSE,
  eopen = TRUE,
  bypass_x_check = FALSE,
  bypass_y_check = FALSE
)

## S4 method for signature 'nanotime,nanotime,nanoperiod,nanoperiod'
align_idx(
  x,
  y,
  start = as.nanoperiod(0),
  end = as.nanoperiod(0),
  sopen = FALSE,
  eopen = TRUE,
  tz,
  bypass_x_check = FALSE,
  bypass_y_check = FALSE
)

## S4 method for signature 'nanotime,nanotime,missing,nanoperiod'
align_idx(
  x,
  y,
  start = as.nanoperiod(0),
  end = as.nanoperiod(0),
  sopen = FALSE,
  eopen = TRUE,
  tz,
  bypass_x_check = FALSE,
  bypass_y_check = FALSE
)

## S4 method for signature 'nanotime,nanotime,nanoperiod,missing'
align_idx(
  x,
  y,
  start = as.nanoperiod(0),
  end = as.nanoperiod(0),
  sopen = FALSE,
  eopen = TRUE,
  tz,
  bypass_x_check = FALSE,
  bypass_y_check = FALSE
)

Arguments

Details

In order to perform the alignment, intervals are created around each elements in y using start and end. For each such interval, the closest element in x is chosen. If no element in x falls in the interval, then NaN is returned.

When only x and y are specified, the default is to close the intervals so that the alignment simply picks up equal points. Note that it is possible to specify meaningless intervals, for instance with a start that is beyond end. In this case, the alignment will simply return NA for each element in y. In principle, the start and end are chosen to define an interval is the past, or around the points in y, but if they are both positive, they can define intervals in the future.

Value

a vector of indices of the same length as y; this vector indexes into x and represent the closest point of x that is in the interval defined around each point in y

Examples

## Not run: 
align_idx(nanotime(c(10:14, 17:19)), nanotime(11:20))
## [1]  2  3  4  5  NA NA  6  7  8  NA

## End(Not run)

Return the number of observations per interval

Description

frequency returns the number of observations in data.table x for each interval specified by by.

Usage

## S4 method for signature 'data.table'
frequency(
  x,
  by,
  grid_start,
  grid_end,
  tz,
  ival_start = -by,
  ival_end,
  ival_sopen = FALSE,
  ival_eopen = TRUE
)

Arguments

Value

a data.table time-series with the number of observations in x that fall withing the intervals defined by the grid interval defined by by.

Examples

## Not run: 
one_second <- as.nanoduration("00:00:01")
one_minute <- 60 * one_second
x <- data.table(index=nanotime((1:100) * one_second), 1)
setkey(x, index)
frequency(x, one_minute)

## End(Not run)

Align a data.table onto a nanotime vector grid

Description

grid_align returns the subset of data.table x that aligns on the grid defined by by, start and end

Usage

grid_align(x, by, ...)

## S4 method for signature 'data.table,nanoduration'
grid_align(
  x,
  by,
  func = NULL,
  grid_start = x[[1]][1] + by,
  grid_end = tail(x[[1]], 1),
  ival_start = -by,
  ival_end = as.nanoduration(0),
  ival_sopen = FALSE,
  ival_eopen = TRUE
)

## S4 method for signature 'data.table,nanoperiod'
grid_align(
  x,
  by,
  func = NULL,
  grid_start = plus(x[[1]][1], by, tz),
  grid_end = tail(x[[1]], 1),
  ival_start = -by,
  ival_end = as.nanoperiod(0),
  ival_sopen = FALSE,
  ival_eopen = TRUE,
  tz
)

Arguments

Details

A grid defined by the parameter by, start and end is created. The function then does a standard alignment of x onto this grid (see the align function)

Value

a data.table time-series of the same length as y with the aggregations computed by func

Examples

## Not run: 
one_second <- 1e9
x <- data.table(index=nanotime(cumsum(sin(seq(0.001, pi, 0.001)) * one_second)))
x <- x[, V2 := 1:nrow(x)]
setkey(x, index)
grid_align(x, as.nanoduration("00:01:00"), sum)

## End(Not run)

Arithmetic operations on two data.table time-series

Description

ops returns the y time-series on which the x time-series values are applied using the specified operator op.

Usage

ops(x, y, op_string)

## S4 method for signature 'data.table,data.table,character'
ops(x, y, op_string)

Arguments

Details

The n elements of the x time-series operand define a set of n-1 intervals, and the value associated with each interval is applied to all the observations in the y time-series operand that fall in the interval. Note that the interval is closed at the beginning and open at the end. The supported values for op are "*", "/", "+", "-".

There has to be one numeric column in x and y; there has to be either a one to one correspondance between the number of numeric columns in x and y, or there must be only one numeric column in x that will be applied to all numeric columns in y. Non-numeric columns must not appear in x, whereas they will be skipped of they appear in y.

Examples

## Not run: 
one_second_duration  <- as.nanoduration("00:00:01")
t1 <- nanotime(1:2 * one_second_duration * 3)
t2 <- nanotime(1:4 * one_second_duration)
dt1 <- data.table(index=t1, data1 = 1:length(t1))
setkey(dt1, index)
dt2 <- data.table(index=t2, data1 = 1:length(t2))
setkey(dt2, index)
ops(dt1, dt2, "+")

## End(Not run)