Skip to contents

Run length encoding to detect consecutive sequences of events.

Usage

rle2(x, index, l_run, value)

Source

Raphael Saldanha, https://rfsaldanha.github.io/posts/run_length_encoding.html with modifications by Adam H. Sparks.

Arguments

x

A vector of values to check for consecutive sequences of events.

index

An index of event values, e.g. a vector of date values.

l_run

An integer value indicating the length of the run for the period of interest.

value

The base value to check against, for which values greater than or equal to value will be recorded as TRUE.

Value

A data.table with three columns, ‘index’, the index values provided; ‘value’, the original values of x that were tested; and ‘test’, a column of Boolean values indicating whether the value was part of a run of at least l_run consecutive values greater than or equal to value (TRUE) or not (FALSE). Missing values in x are never part of a run.

Note

R already has an base::rle(), so this function is suffixed with 2 to avoid NAMESPACE clashes.

Author

Raphael Saldanha and Adam Sparks, adamhsparks@gmail.com

Examples


# Get rainfall data since 2017 for Northam
library(weatherOz)
library(scales)
library(ggplot2)

wd <- get_dpird_summaries(
  station_code = "NO",
  start_date = "20170101",
  end_date = "20171231",
  api_key = Sys.getenv("DPIRD_API_KEY"),
  interval = "daily",
)

# Determine where runs of 2 or more days with rain over 0.5mm occur
rain <- rle2(
  x = wd$rainfall,
  index = wd$date,
  l_run = 2,
  value = 0.5
)

# Plot rainfall values
ggplot(data = rain, aes(x = index, y = value)) +
  geom_line(colour = "light blue", alpha = 0.7) +
  geom_point(aes(color = test), alpha = 0.85) +
  scale_color_manual(values = c("blue", "darkred")) +
  scale_x_date(labels = date_format("%m-%Y")) +
  xlab("Date") +
  ylab("Rainfall (mm)") +
  guides(colour = guide_legend("Two or more consecutive days of rain")) +
  theme(legend.position = "bottom", legend.direction = "horizontal")


# Determine where runs of 2 or more days avg temp above 28 C occur
tavg <- rle2(
  x = wd$air_tavg,
  index = wd$date,
  l_run = 2,
  value = 28
)

# Plot temperature values
ggplot(data = tavg, aes(x = index, y = value)) +
  geom_line(colour = "lightblue", alpha = 0.7) +
  geom_point(aes(color = test), alpha = 0.85) +
  scale_color_manual(values = c("blue", "darkred")) +
  geom_hline(yintercept = 28, colour = "darkred") +
  scale_x_date(labels = date_format("%m-%Y")) +
  xlab("Date") +
  ylab("Temperature (C)") +
  guides(colour = guide_legend("Two or more consecutive days above 28 C")) +
  theme(legend.position = "bottom", legend.direction = "horizontal")