subsecond-tibbletime.R

``` r
library(tibbletime)
library(dplyr)

options(digits.secs = 6)

# 1000 values
ex <- create_series('2013' ~ '2013-01-01 00:00:00.999', period = "1 millisec")

# Random data
ex$numbers <- rnorm(nrow(ex))

ex
#> # A time tibble: 1,000 x 2
#> # Index: date
#>    date                numbers
#>    <dttm>                <dbl>
#>  1 2013-01-01 00:00:00   1.70 
#>  2 2013-01-01 00:00:00  -0.902
#>  3 2013-01-01 00:00:00  -0.199
#>  4 2013-01-01 00:00:00  -2.27 
#>  5 2013-01-01 00:00:00   0.143
#>  6 2013-01-01 00:00:00  -0.485
#>  7 2013-01-01 00:00:00   0.304
#>  8 2013-01-01 00:00:00  -0.101
#>  9 2013-01-01 00:00:00   0.112
#> 10 2013-01-01 00:00:00   2.14 
#> # ... with 990 more rows

ex_mean <- ex %>%
  mutate(date = collapse_index(date, "15 millisec")) %>%
  group_by(date) %>%
  summarise(mean_num = mean(numbers))

ex_mean
#> # A time tibble: 67 x 2
#> # Index: date
#>    date                mean_num
#>    <dttm>                 <dbl>
#>  1 2013-01-01 00:00:00  -0.0912
#>  2 2013-01-01 00:00:00   0.0876
#>  3 2013-01-01 00:00:00  -0.306 
#>  4 2013-01-01 00:00:00  -0.211 
#>  5 2013-01-01 00:00:00  -0.0460
#>  6 2013-01-01 00:00:00   0.366 
#>  7 2013-01-01 00:00:00  -0.241 
#>  8 2013-01-01 00:00:00   0.345 
#>  9 2013-01-01 00:00:00  -0.0739
#> 10 2013-01-01 00:00:00  -0.0801
#> # ... with 57 more rows

# Check against xts
library(xts)

ex_xts <- xts(ex$numbers, ex$date)
ex_xts_mean <- period.apply(ex_xts, endpoints(ex_xts, "milliseconds", 15), mean)

head(ex_xts_mean)
#> Warning: timezone of object (UTC) is different than current timezone ().
#>                                [,1]
#> 2013-01-01 00:00:00.013 -0.09121474
#> 2013-01-01 00:00:00.029  0.08757511
#> 2013-01-01 00:00:00.043 -0.30586630
#> 2013-01-01 00:00:00.059 -0.21118297
#> 2013-01-01 00:00:00.073 -0.04603550
#> 2013-01-01 00:00:00.088  0.36589842

identical(as.vector(coredata(ex_xts_mean)), ex_mean$mean_num)
#> [1] TRUE
```