时间戳

首页 > 解决方案 > R Automatically name results of Loop

r - R Automatically name results of Loop

问题描述

I have a list of dataframes:

df_DJF = data.frame(replicate(2,sample(0:130,30,rep=TRUE)))
df_JJA = data.frame(replicate(2,sample(0:130,20,rep=TRUE)))
df_MAM = data.frame(replicate(2,sample(0:130,25,rep=TRUE)))
df_SON = data.frame(replicate(2,sample(0:130,15,rep=TRUE)))

df_list = list(df_DJF, df_JJA, df_MAM, df_SON)

I want to randomly choose 80% of each the dataframe. I can do that manually by doing this and using the sample_size as row index. 

sample_size = floor(0.8*nrow(df_DJF))
picked_DJF = sample(seq_len(nrow(df_DJF)), size = sample_size)

My problem is that I have very many df with different number of rows. So I want to automatize this process. In the end I want to have 4 sample sizes with the correct number in it. The names of the sample_sizes should be:

samplenames = paste("sample_size", c("DJF", "JJA", "MAM", "SON"), sep = "_")

Same for the "picked"...it should be picked_DJF and so on...

标签： rloopssample

解决方案

继续使用列表，而不是assign. 设置您的names(df_list) = c("DJF", "JJA", "MAM", "SON"), 然后为后续列表使用相同的名称，例如picked列表。

# for a single sample size
picked = lapply(df_list, function(x) x[sample(1:nrow(x), size = floor(0.8 * nrow(x))), ])

Usinglapply将保留原始列表的名称，因此您不必担心。

对于来自每个数据帧的多个样本大小，您可以创建一个带有嵌套的嵌套列表lapply：

names(df_list) = c("DJF", "JJA", "MAM", "SON")
sample_prop = list(s1 = 0.2, s2 = 0.4, s3 = 0.6, s4 = 0.8)
picked = lapply(df_list, function(df) lapply(sample_prop, function(sp) {
  df[sample(nrow(df), size = floor(sp * nrow(df))), ]
}))

# then access individual data frames with `$` or `[[`
picked$JJA$s3
#    X1  X2
# 17 70 128
# 7  94 121
# 1  57 125
# 8  32  75
# 9  15   8
# 19 58  15
# 20 55  17
# 10 42  15
# 4  51  67
# 12 89  13
# 2  74  50
# 14 77  36

将数据框划分为“picked”和“unpicked”split是有道理的。它已经返回一个list. 这将给出一个三重嵌套列表结果：

result = lapply(df_list, function(df) lapply(sample_prop, function(sp) {
  n_pick = floor(sp * nrow(df))
  n_unpick = nrow(df) - n_pick
  split(df, f = c(rep("picked", n_pick), rep("unpicked", n_unpick))[sample(nrow(df))])
}))

result$JJA$s3$unpicked
#     X1  X2
# 2   74  50
# 3   62  78
# 4   51  67
# 6  103  42
# 7   94 121
# 11  59  60
# 14  77  36
# 16  83  72

推荐阅读