Geek Answers Handbook

Effectively replace a fixed-position substring with a string of equal or greater length

What would be an efficient way to replace a substring with a fixed position with another string of equal or greater length?

For example, the following submenu “abc” replaces the position “abc” first and then replaces it:

sub("abc", "123", "iabc.def", fixed = TRUE)
#[1] "i123.def"

sub("abc", "1234", "iabc.def", fixed = TRUE)
#[1] "i1234.def"

However, we know that the substring "abc" is always at the positions of the characters 2, 3, and 4. In this case, is there a way to specify these positions so that the string does not match "t need to be done, but use character indices instead?"

I tried using substr (), but it did not work as I had hoped when the replacement string is larger than the substring to replace:

x <- "iabc.def"
substr(x, 2, 4) <- "123"
#[1] "i123.def"

x <- "iabc.def"
substr(x, 2, 4) <- "1234"
#[1] "i123.def"

Thanks so much in advance for your time,

Tony braial

P.S. , , , , :)

===== =====

#                             test elapsed  relative
# 7 francois.fx_wb(x, replacement)    0.94  1.000000
# 1                           f(x)    1.56  1.659574
# 6    francois.fx(x, replacement)    2.23  2.372340
# 5                      Sobala(x)    3.89  4.138298
# 2                    Hong.Ooi(x)    4.41  4.691489
# 3                        DWin(x)    5.57  5.925532
# 4                      hadley(x)    9.47 10.074468

:

library(rbenchmark)
library(stringr)
library(Rcpp)
library(inline)

f <- function(x, replacement = "1234")  sub("abc", replacement, x, fixed = TRUE)

Hong.Ooi <- function(x, replacement = "1234") paste(substr(x, 1, 1), replacement, substr(x, 5, nchar(x)), sep = "")

DWin <- function(x, replacement =  paste("\\1", "1234", sep = "")) sub("^(.)abc", replacement, x)

Sobala <- function(x, replacement =  paste("\\1", "1234", sep = ""))  sub("^(.).{3}", replacement, x, perl=TRUE)

hadley <- function(x, replacement = "1234") {
  str_sub(x, 2, 4) <- replacement
  return(x)
}

francois.fx <- cxxfunction( signature( x_ = "character", rep_ = "character" ), '

    const char* rep =as<const char*>(rep_) ;
    CharacterVector x(x_) ;
    int nrep = strlen( rep ) ;
    int n = x.size() ; 
    CharacterVector res(n) ;

    char buffer[1024] ;

    for(int i=0; i<n; i++) {
        const char* xi = x[i] ;
        if( strncmp( xi+1, "abc", 3 ) ) {
            res[i] = x[i] ;
        } else{
            buffer[0] = xi[0] ;
            strncpy( buffer + 1, rep, nrep ) ;
            strcpy( buffer + 1 + nrep, xi + 4 ) ;
            res[i] = buffer ;
        }
    }
    return res ;
', plugin = "Rcpp" )

francois.fx_wb <- cxxfunction( signature( x_ = "character", rep_ = "character" ), '

    const char* rep =as<const char*>(rep_) ;
    int nrep = strlen( rep ) ;
    int n=Rf_length(x_) ;
    SEXP res = PROTECT( Rf_allocVector( STRSXP, n ) ) ;

    char buffer[1024] ;

    for(int i=0; i<n; i++) {
        const char* xi = char_get_string_elt(x_, i) ;
        if( strncmp( xi+1, "abc", 3 ) ) {
            set_string_elt( res, i, get_string_elt(x_,i)) ;
        } else{
            buffer[0] = xi[0] ;
            strncpy( buffer + 1, rep, nrep ) ;
            strcpy( buffer + 1 + nrep, xi + 4 ) ;
            char_set_string_elt(res, i, buffer ) ;
        }
    }
    UNPROTECT(1) ;
    return res ;
', plugin = "Rcpp" )


x <- rep("iabc.def", 1e6)
replacement <- "1234"
benchmark(f(x), Hong.Ooi(x), DWin(x), hadley(x), Sobala(x), francois.fx(x, replacement), francois.fx_wb(x, replacement),
          columns = c("test", "elapsed", "relative"),
          order = "relative",
          replications = 10)
+5
4

Rcpp.

fx <- cxxfunction( signature( x_ = "character", rep_ = "character" ), '

    const char* rep =as<const char*>(rep_) ;
    CharacterVector x(x_) ;
    int nrep = strlen( rep ) ;
    int n = x.size() ; 
    CharacterVector res(n) ;

    char buffer[1024] ;

    for(int i=0; i<n; i++) {
        const char* xi = x[i] ;
        if( strncmp( xi+1, "abc", 3 ) ) {
            res[i] = x[i] ;
        } else{
            buffer[0] = xi[0] ;
            strncpy( buffer + 1, rep, nrep ) ;
            strcpy( buffer + 1 + nrep, xi + 4 ) ;
            res[i] = buffer ;
        }
    }
    return res ;
', plugin = "Rcpp" )

, R . , , , , , :/

fx_wb <- cxxfunction( signature( x_ = "character", rep_ = "character" ), '

    const char* rep =as<const char*>(rep_) ;
    int nrep = strlen( rep ) ;
    int n=Rf_length(x_) ;
    SEXP res = PROTECT( Rf_allocVector( STRSXP, n ) ) ;

    char buffer[1024] ;

    for(int i=0; i<n; i++) {
        const char* xi = char_get_string_elt(x_, i) ;
        if( strncmp( xi+1, "abc", 3 ) ) {
            set_string_elt( res, i, get_string_elt(x_,i)) ;
        } else{
            buffer[0] = xi[0] ;
            strncpy( buffer + 1, rep, nrep ) ;
            strcpy( buffer + 1 + nrep, xi + 4 ) ;
            char_set_string_elt(res, i, buffer ) ;
        }
    }
    UNPROTECT(1) ;
    return res ;
', plugin = "Rcpp" )

R Internals :

" ", - ( STRSXP). , , , . SET_VECTOR_ELT SET_STRING_ELT, . , SEXP ( lvalues ​​ C).

R- .

, :

. , . , , , . , , node .

+2

:

> sub("^(.)abc", "\\1xyz", c("aabcdef", "xxxxxxx"))
[1] "axyzdef" "xxxxxxx"
+2

The easiest way I can think of:

x <- paste(substr(x, 1, 1), "1234", substr(x, 5, nchar(x)), sep="")
+1
source

Some improvement in the DWin function.

function(x, replacement =  paste("\\1", "1234", sep = "")) 
                     sub("^(.).{3}", replacement, x,perl=TRUE)
+1
source

More articles:


All Articles