Extract digits with decimal from a vector R-CodePudding

I have the following vector:

c("c(`Kruskal-Wallis chi-squared` = 201.760850624131)", "c(df = 17)", 
"1.26686891197831e-33", "Kruskal-Wallis rank sum test", "delta_Z by criteria"
)

I desired this output:

c("201.760850624131", "17", "1.26686891197831e-33")

Thanks for any help

CodePudding user response：

We can try using str_extract here for a regex option:

x <- c("c(`Kruskal-Wallis chi-squared` = 201.760850624131)",
       "c(df = 17)", 
       "1.26686891197831e-33",
       "Kruskal-Wallis rank sum test",
       "delta_Z by criteria"
)
output <- as.numeric(na.omit(str_extract(x, "\\d (?:\\.\\d )?(?:e[ -]\\d )?")))
output

[1] 2.017609e 02 1.700000e 01 1.266869e-33

Explanation of regex:

\\d match an integer component
(?:\\.\\d )? optional decimal component
(?:e[ -]\\d )? optional exponent (either positive or negative)

CodePudding user response：

We can use str_extract with the regex that matches one or more digits ([0-9] ) followed by a . then one or more digits and e followed by - or and any digits

library(stringr)
as.numeric(na.omit(str_extract(v1, "[0-9] (\\.[0-9] e[- ]\\d )?")))

-output

[1] 2.010000e 02 1.700000e 01 1.266869e-33

data

v1 <- c("c(`Kruskal-Wallis chi-squared` = 201.760850624131)", "c(df = 17)", 
"1.26686891197831e-33", "Kruskal-Wallis rank sum test", "delta_Z by criteria"
)

CodePudding user response：

Since it appears to be R code:

x <- c("c(`Kruskal-Wallis chi-squared` = 201.760850624131)", "c(df = 17)", 
"1.26686891197831e-33", "Kruskal-Wallis rank sum test", "delta_Z by criteria")

as.character(sapply(sapply(x[1:3], str2lang, USE.NAMES = F), eval))

#> [1] "201.760850624131"     "17"                   "1.26686891197831e-33"