In ijlyttle/shinychord: Reusable shiny modules

library("knitr")
library("lubridate")
library("magrittr")

opts_chunk$set(comment = "", echo = FALSE)

fn_parse <- function(x, tz_parse, tz_display){

  dtm <- lubridate::ymd_hms(x, tz = tz_parse) 
  dtm <- lubridate::with_tz(dtm, tzone = tz_display)

  dtm
}

format_tz <- function(x){
  paste0(format(x), " (", lubridate::tz(x), ")")
}

str_time <- "2015-09-27 21:45:00"

tz_parse <- "America/Chicago"
tz_display <- "Europe/Paris"

Short answer

• Timezone to parse: describes the timezone used to write to the file. The timestamps and the parsing timezone define the instants in time.
• Timezone to display: This is used to specify the display of these instants in time, but will not change the value of these instants.

The default behavior of this app is to set the display timezone to value of the parsing timezone.

Long answer

There are two steps to processing text-based timestamps:

1. Determining the instants in time to which the timestamps refer.
2. Displaying those instants in time the way that you wish.

Let's say your friend tells you that when she saw the peak of the "blood moon" eclipse, the clock read:

str_time %>% cat()

Knowing only what the clock read is not enough information to determine the instant in time when the eclipse happened; you need to know where the clock was.

In this case, let's say your friend was in Chicago. Now we have enough information to fix the instant in time. We can parse the timestamp to describe the instant, by referring to the timezone by the place name "r tz_parse".

str_time %>% fn_parse(tz_parse, tz_parse) %>% format_tz() %>% cat()

Timezones are named either "UTC" (also known as GMT, or Zulu) or after locations inside each particular timezone, for example "America/Chicago", rather than the more-familiar "Central Time", or "US Central Time".

Let's say that you are in Paris, and you would like to see what your clock would display at this instant in time. We would then apply the timezone reference named "r tz_display":

str_time %>% fn_parse(tz_parse, tz_display) %>% format_tz() %>% cat()

In order to parse and display the instants of time, we use these two separate contexts for the timezone.

Please note: normally, the timezone will not be printed - we do so here to help illustrate the concepts. More examples are provided to demonstrate some different cases.

Example 1

tz_parse <- "Europe/London"
tz_display <- tz_parse

str_time <- "2015-01-02 03:00:00"

Your file describes the energy usage at a facility in the United Kingdom. The timestamps are written using the local (UK) timezone; you wish to display the time in the local timezone.

In the file, one of the timestamps may be written as:

str_time %>% cat()

In this case, use:

• Timezone to parse: "r tz_parse"
• Timezone to display: "r tz_display"

The parsed timestamp will be displayed as:

str_time %>% fn_parse(tz_parse, tz_display) %>% format_tz() %>% cat()

Example 2

tz_parse <- "UTC"
tz_display <- "Asia/Seoul"

str_time <- "2015-04-05 06:00:00"

Your file describes weather observations in South Korea. The timestamps are written using UTC; you wish to display the time in the local (South Korean) timezone.

In the file, one of the timestamps may be written as:

str_time %>% cat()

In this case, use:

• Timezone to parse: "r tz_parse"
• Timezone to display: "r tz_display"

The parsed timestamp will display as:

str_time %>% fn_parse(tz_parse, tz_display) %>% format_tz() %>% cat()

Example 3

tz_parse <- "UTC"
tz_display <- "America/Chicago"

str_time <- "2015-07-08T09:00:00-0500"

Your file describes the energy-generation rate of a wind farm in Iowa (USA). The timestamps are written using ISO-8601 format; you wish to display the time in the local (Iowa) timezone. The closest major center in the same timezone is Chicago.

In the file, one of the timestamps may be written as:

str_time %>% cat()

In this case, use:

• Timezone to parse: does not matter
• Timezone to display: "r tz_display"

The parsed timestamp will display as:

str_time %>% fn_parse(tz_parse, tz_display) %>% format_tz() %>% cat()

You may wonder why it does not matter what you specify as the parsing timezone. The ISO-8601 format is designed such that the timestamp is always parsed using "UTC". In this case, it will not matter what we specify as the parsing timezone; only the display timezone is operative.

ijlyttle/shinychord documentation built on May 18, 2019, 3:41 a.m.