在处理因素时，Base R grep家族比stringr变体快得多

搜索唯一值比搜索所有元素更快

正如MrFlick在 comments 中指出的那样，这种情况的发生是因为因子是grepl()中的一种特例，它匹配水平，而不是整个载体.您有一个包含一百万个元素但只有三个唯一值的载体，所以这会减少很多工作量.

我认为stringr没有理由告诉它你正在处理具有许多重复值的数据.但是，您可以编写一个将grepl()逻辑应用于stringr的因子的函数:

str_detect_factor <- function(fct, pattern, negate = FALSE) {
    out <- stringr::str_detect(
        c(levels(fct), NA_character_), pattern, negate
    )
    outna <- out[length(out)]
    out <- out[fct]
    out[is.na(fct)] <- outna
    out
}

这比grepl()稍微快一点:

bench::mark(
    base_facts = grepl("Two", sample_facts),
    stringr_facts = str_detect(sample_facts, "Two"),
    stringr_facts2 = str_detect_factor(sample_facts, "Two")
)

#   expression          min   median `itr/sec` mem_alloc `gc/sec` n_itr  n_gc total_time
#   <bch:expr>     <bch:tm> <bch:tm>     <dbl> <bch:byt>    <dbl> <int> <dbl>   <bch:tm>
# 1 base_facts       3.27ms   4.15ms    220.      11.4MB    112.     49    25      222ms
# 2 stringr_facts   182.1ms 182.44ms      5.44    11.4MB      0       3     0      552ms
# 3 stringr_facts2   3.33ms   3.75ms    252.      11.4MB     96.6    73    28      290ms

将此优化应用于字符载体

您的字符载体也有许多重复的值，因此您可以应用相同的逻辑.这不会像因子那么快，因为不同的值已经存储在对象中，所以您不必调用unique().但是，它应该比现有方法更快:

str_detect_rep <- function(string, pattern, negate = FALSE) {
    unique_string <- c(unique(string), NA_character_)
    out <- stringr::str_detect(
        unique_string, pattern, negate
    ) |> setNames(unique_string)
    outna <- out[length(out)]
    out <- out[string]
    out[is.na(string)] <- outna
    unname(out)
}

我们可以看到它比基本R或stringr版本快得多:

bench::mark(
    base_chars = grepl("Two", sample_chars),
    stringr_chars = str_detect(sample_chars, "Two"),
    stringr_chars2 = str_detect_rep(sample_chars, "Two")
)

# A tibble: 3 × 9
#   expression          min   median `itr/sec` mem_alloc `gc/sec` n_itr  n_gc total_time
#   <bch:expr>     <bch:tm> <bch:tm>     <dbl> <bch:byt>    <dbl> <int> <dbl>   <bch:tm>
# 1 base_chars      131.2ms  173.1ms      5.80    3.81MB     0        4     0    689.6ms
# 2 stringr_chars     213ms  217.5ms      4.60    3.81MB     2.30     2     1    434.9ms
# 3 stringr_chars2   37.3ms   37.3ms     26.8    42.33MB   295.       1    11     37.3ms

创建常规

由于您的问题是您不想考虑是否正在处理字符数据或因子数据，因此您可以创建一个通用:

str_detect2 <- function(string, ...) UseMethod("str_detect2")
str_detect2.default <- function(string, ...) str_detect_rep(string, ...)
str_detect2.factor <- function(string, ...) str_detect_factor(string, ...)

然后进行基准测试:

bench::mark(
    base_facts = grepl("Two", sample_facts),
    stringr_facts = str_detect(sample_facts, "Two"),
    str_detect2_facts = str_detect2(sample_facts, "Two"),
    base_chars = grepl("Two", sample_chars),
    stringr_chars = str_detect(sample_chars, "Two"),
    str_detect2_chars = str_detect2(sample_chars, "Two")
)

#   expression             min   median `itr/sec` mem_alloc `gc/sec` n_itr  n_gc total_time
#   <bch:expr>        <bch:tm> <bch:tm>     <dbl> <bch:byt>    <dbl> <int> <dbl>   <bch:tm>
# 1 base_facts          3.77ms    6.3ms    154.     11.44MB     40.3    42    11      273ms
# 2 stringr_facts     197.82ms  207.1ms      4.65   11.44MB      0       3     0      645ms
# 3 str_detect2_facts   3.55ms    4.8ms    190.     11.44MB     43.5    70    16      368ms
# 4 base_chars        121.17ms  125.8ms      7.47    3.81MB      0       4     0      536ms
# 5 stringr_chars     215.54ms  232.9ms      4.36    3.81MB      0       3     0      689ms
# 6 str_detect2_chars   27.5ms   29.2ms     31.2    42.33MB     85.8     4    11      128ms

现在您只需拨打str_detect2()而不必考虑它-字符和因子数据这里速度更快: