ggplot2中y比例标签的好K/M/G缩写

Posted 2023-03-07

【中文标题】ggplot2中y比例标签的好K/M/G缩写

【英文标题】：Good K/M/G abbreviation of y scale labels in ggplot2

【发布时间】：2019-12-04 09:22:40

【问题描述】：

问题

我们如何轻松拥有没有“位”或“字节”单位符号的 Kilo/Mega/Giga 标签？

示例

data.frame(x = LETTERS[1:5], n = c(0, 5000, 10000, 15000, 20000)) %>% 
  ggplot(aes(x, n)) + 
  geom_point() +
  scale_y_continuous(labels = scales::number_bytes_format(units = "si"))

对于 y 比例，我希望标签为 0K、5K、10K、15K、20K。没有Kb！

奖金问题

是否有任何可用的解决方案来获得0、1K、1M、1G 标签？即，最合适的值缩写？

【问题讨论】：

替代，除以 1000，并将 y 标记为“n (Kb)”？

我不知道规模的大小。 Il 可能是 K，可能是 M 甚至 G。到现在为止，我只想摆脱 "Kb" 中的 "b"

相关帖子：***.com/q/29787452/680068

来自相关帖子：gdata::humanReadable(n, standard = "Unix", sep = "")

【参考方案1】：

试试gdata::humanReadable:

library(ggplot2)
library(gdata)

myDat <- data.frame(x = LETTERS[1:5], n = c(0, 5000, 10000, 15000, 20000))

ggplot(myDat, aes(x, n)) + 
  geom_point() +
  scale_y_continuous(breaks = myDat$n, 
                     labels = humanReadable(myDat$n, standard = "Unix", sep = ""))

---

编辑：

我们可以自定义函数：

humanReadableCustom <- function (x, units = "auto", standard = c("IEC", "SI", "Unix"), 
                                 digits = 1, width = NULL, sep = " ", justify = c("right", 
                                                                                  "left")) 

  #suffix.SI <- c("B", "kB", "MB", "GB", "TB", "PB", "EB", "ZB", "YB")
  # custom
  suffix.SI <- c("", "K", "M", "G", "T", "P", "E", "Z", "Y")

  suffix.IEC <- c("B", "KiB", "MiB", "GiB", "TiB", "PiB", "EiB", "ZiB", "YiB")
  suffix.Unix <- c("B", "K", "M", "G", "T", "P", "E", "Z", "Y")
  standard <- match.arg(standard)
  if (length(justify) == 1) 
    justify <- c(justify, justify)
  .applyHuman <- function(x, base, suffix, digits, width, 
                          sep) 
    n <- length(suffix)
    i <- pmax(pmin(floor(log(x, base)), n - 1), 0)
    if (!is.finite(i)) 
      i <- 0
    x <- x/base^i
    if (is.null(width)) 
      x <- format(round(x = x, digits = digits), nsmall = digits)
    else 
      lenX <- nchar(x)
      if (lenX > width) 
        digits <- pmax(width - nchar(round(x)) - 1, 
                       0)
      
      if (i == 0) 
        digits <- 0
      x <- round(x, digits = digits)
    
    c(x, suffix[i + 1])
  
  if (any(x < 0)) 
    stop("'x' must be positive")
  if (standard == "SI") 
    suffix <- suffix.SI
    base <- 10^3
  
  else if (standard == "IEC") 
    suffix <- suffix.IEC
    base <- 2^10
  
  else 
    suffix <- suffix.Unix
    base <- 2^10
  
  if (!missing(units) && units == "bytes") 
    retval <- rbind(x, "bytes")
  
  else if (!missing(units) && units != "auto") 
    units <- suffix[match(toupper(units), toupper(suffix))]
    power <- match(units, suffix) - 1
    X <- x/(base^power)
    X <- format.default(x = X, digits = digits, nsmall = digits)
    retval <- rbind(X, rep(units, length(X)))
  
  else retval <- sapply(X = x, FUN = ".applyHuman", base = base, 
                        suffix = suffix, digits = digits, width = width, sep = sep)
  if (all(justify == "none")) 
    paste(trim(retval[1, ]), trim(retval[2, ]), sep = sep)
  else paste(format(trim(retval[1, ]), justify = justify[1]), 
             format(trim(retval[2, ]), justify = justify[2]), sep = sep)

然后绘制

library(ggplot2)
library(gdata)

myDat <- data.frame(x = LETTERS[1:5], n = c(0, 5000, 10000, 15000, 20000))

ggplot(myDat, aes(x, n)) + 
  geom_point() +
  scale_y_continuous(breaks = myDat$n, 
                     labels = humanReadableCustom(myDat$n,
                                                  standard = "SI", sep = ""))

【讨论】：

谢谢，但这不是我想要的 • standard : character, "IEC" for powers of 1024 ('MiB'), "SI" for powers of 1000 ('MB'), or "Unix" for powers of 1024 ('M'). See details. • 我想要 1000（'M'）的幂，并且可能使用 g（克）、m（米）作为单位而不是 b （位）

@Costin 查看编辑，我们可以窃取代码并根据需要进行更新。

如果我们偷窃，我更喜欢从这里：***.com/a/1094933/2107667 :) 谢谢@zx8754

@Costin 没问题，这只是一个想法，你可以随意定制。