match_col: Match columns to list elements

In pttry/statficlassifications: Statistics Finland Statistical Classification API interface

Match columns to list elements

Description

Find columns in list that match vector by values or by names. Used to automatically detect the columns in data.frame that should be recoded given a key.

Usage

match_col(x, y, x_name = deparse(substitute(x)), by = "names")

Arguments

x	vector, factor or data.frame.
x_name	name of input (column) to tell which is the from column in key.
by	"names" or "values" whether to match columns by their names or by the values they contain.
df	data.frame where matcheds are found.

Details

For each column of x, find the columns in y that has a variable with the same classification as the column is x has.

by_values: find the column in y that contains most elements of (column of) x to map the coding of (column of) x to (a column of) y that has the same coding.

by_names: find the column in y that has the same name as (a column of) x to map the coding of (column of) x to (a column of) y that has the same coding.

Value

list with an element for each column in x. Each element of this list contains all a vector with the names of columns that where matched in y.

Examples

v <- c("a", "b", "a", "c", "b", "b")
f <- factor(v)
df <- data.frame(var1 = v, y = rnorm(6))

key <- data.frame(var1 = letters[1:4],
                  var2 = c("first letter",
                           "second letter",
                           "third letter",
                           "fourth letter"))

 match_col(v, key, by = "values")
 var1 <- v
 match_col(var1, key)
 match_col(f, key, by = "values")
 match_col(df, key)
 match_col(df, key, by = "values")

key <- data.frame(var1 = letters[1:4],
                  var3 = letters[4:1],
                 var2 = c("first letter",
                          "second letter",
                          "third letter",
                          "fourth letter"))
 match_col(v, key, by = "values")
 match_col(df, key, by = "names")
 match_col(df, key, by = "values")


 # Given the iris data set. Find in that data set the column that contains
 # the same values as vector v and thus likely describes the same variable
 data(iris)
 v <- c("setosa", "versicolor")
 match_col(v, iris, by = "values")

 # Now suppse there is a key that maps the Latin Species name to a common
 # name
 key <- data.frame(Species = c("setosa", "versicolor", "virginica"),
                   common_name = c("bristle-pointed iris",
                            "harlequin blueflag", "Virginia blueflag"))

 # We find for each column in iris, a column in the key that has data on the
 # same variable as in the column. Here by names:
 match_col(iris, key, by = "names")

 # But say the key has a different variable name even though the variable is
 # the same:
 names(key)[1] <- "Genus"
 match_col(iris, key, by = "values")

 # Sometime's inputs do not have clear names to use in mapping to key
 match_col(iris$Species, key, by = "values")

pttry/statficlassifications documentation built on Jan. 17, 2024, 4:36 p.m.