str_pad count 1 chinese char as 2

[frootor]

vchs= "中,中中,中中中"
yc=str_split(vchs,",",simplify = T)
print(yc)
[,1] [,2] [,3]
[1,] "中" "中中" "中中中"
yl=str_length(yc)
print(yl)
[1] 1 2 3
print(max(yl))
[1] 3
ycd=stringr::str_pad(yc,width=max(str_length(yc)),side = "right",pad=" ")
ycd
[1] "中 " "中中" "中中中"
ycd=stringr::str_pad(yc,width=2*max(str_length(yc)),side = "right",pad=" ")
ycd
[1] "中 " "中中 " "中中中"

[frootor]

str_pad<- function(string, width, side = c("left", "right", "both"), pad = " ") {
side <- match.arg(side)
switch(side,
left = stri_pad_left(string, width, pad = pad, use_length = T ),
right = stri_pad_right(string, width, pad = pad, use_length = T ),
both = stri_pad_both(string, width, pad = pad, use_length = T )
)
}

[yutannihilation]

Hi, let me comment on this as this feature is added by my request on stringi's side (gagolews/stringi#149).

I believe it is fine to pad based on their widths most of the time, assuming the main use case of padding is to arrange strings prettier with the proper number of white spaces:

reprex::reprex_info()
#> Created by the reprex package v0.1.1.9000 on 2017-11-06

library(stringr)

c("中", "中中", "中中中") %>%
  str_pad(width = 10) %>%
  cat(sep = "\n")
#>         中
#>       中中
#>     中中中

That said, it may be useful to specify use_length from stringr's function for some cases. I will create a PR ;)

[hadley]

@frootor can you clarify why you need this?

[frootor]

in chinese,there is a full width（全角） space "　" and a half width（半角） space " "。
if i want to pad chinese string to the same width,using a full width sapce "　" is a better choice, for in this case the width and the length are both the same. if use a half width space " " to pad the same width, the strings length are not the same ,for 2 half width space equal the length of a chinese char.
so for a simple logic to handle such problem , add a "use_length" param maybe better.

[frootor]

when using Rgraphviz,i want to pad the node label to the same length and the same width, but str_pad fail to handle it in default way ,then i found this issue.

[frootor]

if not add "use_length", when use str_pad to handle chinese string ,chinese char should count 2 ,then str_pad works, but usually we count a chinese char as 1.

[yutannihilation]

if use a half width space " " to pad the same width, the strings length are not the same ,for 2 half width space equal the length of a chinese char.

Ah, this is true if the font is not monospace.

Still, I wonder the option is really useful for you. You cannot use full-width whitespace "　"(U+3000, IDEOGRAPHIC SPACE) for padding. How will you use use_length to solve your problem?

stringr::str_pad(c("中", "中中", "中中中"), width = 6, pad = "\u3000")
#> Error in stri_pad_left(string, width, pad = pad, use_length = use_length) : 
#>   each string in `pad` should consist of code points of total width 1

I think you can either:

1. use some monospace font (e.g. Noto Sans Mono CJK) or
2. pad with half-width spaces and replace two half-width spaces with one full-width space:

c("中", "中中", "中中中") %>%
  str_pad(width = 6, pad = " ") %>%
  str_replace_all("  ", "\u3000")

[yutannihilation]

Sorry! I misunderstood about what you are saying...

use_length=TRUE must be specified if we want to pad with "wide" characters.

stringr::str_pad(c("中", "中中", "中中中"), width = 6, pad = "\u3000")
#> Error in stri_pad_left(string, width, pad = pad, use_length = use_length) : 
#>   each string in `pad` should consist of code points of total width 1

stringi::stri_pad_left(c("中", "中中", "中中中"), width = 3, pad = "\u3000", use_length = TRUE)
#> [1] "　　中" "　中中" "中中中"

It seems reasonable to add this option on stringr's side.

[frootor]

yes， i finally use stringi::stri_pad settle it