meta_apply: Transforms data based on an other meta table

In Ostluft/rOstluft: Tools for handling air quality data by Ostluft

Transforms data based on an other meta table

Description

To work with data from different sources a normilization of naming is necessary. This function allows flexible renaming/recoding of the input data based on a second table. In short following operation is performed:

data$data_dest = meta[meta$meta_key == data$data_src]$meta_val

For handling missing values for meta_key in meta there are different modes:

• strict: stop execution

• drop: drop rows from data

• keep: keep the values in data

• replace: use the mapping provided in the argument replacement (named vector/list)

The function is quiet chatty and reports which values are dropped, kept or replaced.

Usage

meta_apply(
  data,
  meta,
  data_src,
  data_dest,
  meta_key,
  meta_val,
  mode = "strict",
  replacements = NULL
)

Arguments

data	Input data as tibble
meta	Lookup table as tibble
data_src	Name of the column in data used to lookup in meta
data_dest	Name of the column to save the result in data
meta_key	Name of the column in meta to match against data$data_src
meta_val	Name of the column containing the replacement value in meta
mode	One of "strict", "drop", "keep", "replace". Default "strict"
replacements	Named vector/list with missing values in meta$meta_key or to overwrite specific mappings

Value

transformed data

See Also

• Under the hood the heavy lifting is done by dplyr::recode().

Examples

meta_fn <- system.file("extdata", "meta_smn.rds",
                       package = "rOstluft.data", mustWork = TRUE)
meta <- readRDS(meta_fn)
tibble::glimpse(meta)

fn <- system.file("extdata", "smn.txt", package = "rOstluft.data", mustWork = TRUE)
data <- read_smn(fn, na.rm = FALSE)
data <- dplyr::arrange(data, .data$starttime)
data

# data contains no units, cryptic SwissMetNet parameter names and abbreviations for site.
# And the meta data for parameter rre150z0 is missing. Perfect!

# too lazy to update meta, add unit mapping based on SwissMetNet parameter names
# and we want to overwrite the mapping for dkl010z0 anyway => use replace
res <- meta_apply(data, meta, "parameter", "unit", "parameter_original", "unit",
         mode = "replace",replacements = list(rre150z0 = "unit1", dkl010z0 = "unit2"))
res

# rename the SwissMetNet Parameters, still no mapping for rre150z0, we aren't
# interested in the data and drop it
res2 <- meta_apply(res, meta, "parameter", "parameter",
         "parameter_original", "parameter", mode = "drop")
res2

# or we keep it
res <- meta_apply(res, meta, "parameter", "parameter",
         "parameter_original", "parameter", mode = "keep")
res

# rename the site abbreviation to the site name, strict should work
res <- meta_apply(res, meta, "site", "site", "site_short", "site")
res

Ostluft/rOstluft documentation built on Feb. 6, 2024, 1:26 a.m.