运行R抽样函数时遭遇sample.int无效首参数错误求助

我运行一个抽样函数时反复遇到无法定位的错误，错误信息如下：

Error in sample.int(length(x), size, replace, prob) :
invalid first argument

我的代码如下：

idpSample <- function(x) {

main.sample <- sample(
idp.frame[COUNTY == idp.breakdown[x, County], GEOID9],
idp.breakdown[x, idp_cluster],
prob = idp.frame[COUNTY == idp.breakdown[x, County], mt12_est_idps_hh],
replace = T 
)

 n.reserve <- min(3, ifelse(
 nrow(idp.frame[COUNTY == idp.breakdown[x, County]]) - idp.breakdown[x, idp_cluster] <
  0,0,
nrow(idp.frame[COUNTY == idp.breakdown[x, County]]) - idp.breakdown[x, idp_cluster]
))
if(n.reserve>0){
reserve.sample <- sample(
  idp.frame[COUNTY == idp.breakdown[x, County] & GEOID9 %nin% main.sample, GEOID9],
  n.reserve,
  prob = idp.frame[COUNTY == idp.breakdown[x, County] & GEOID9 %nin% main.sample, 
  mt12_est_idps_hh],
  replace = F 
)
}else{
reserve.sample <- c()
 }

 data.frame(
 GEOID9 = c(main.sample, reserve.sample),
 sample_type = c(rep("Main sample", idp.breakdown[x, idp_cluster]),
                rep("Reserve",  n.reserve)))
}

idp.sample <- rbindlist(lapply(1:nrow(idp.breakdown), idpSample))

样本数据框 #

idp.breakdown 前几行 #

State    County idp_cluster idp_n ret_cluster ret_n hc_cluster hc_n total_cluster total_n
1: Abyei Administrative Area     Abyei          11   143          11   143         11  143            33     429
2:         Central Equatoria      Juba           5    65           3    39          4   52            12     156
3:         Central Equatoria Kajo-Keji           1    13           3    39          8  104            12     156
4:         Central Equatoria    Lainya           4    52           2    26          4   52            10     130
5:         Central Equatoria    Morobo           5    65           3    39          3   39            11     143
6:         Central Equatoria  Terekeka           1    13           1    13         10  130            12     156

idp.frame 前几行 #

STA_CODE      STATE CTY_CODE COUNTY PAY_CODE BOM_CODE FSNMS_Urban mt12_est_idps_ind mt12_est_idps_hh
1:       71 Upper Nile       01   Renk       01      001       FALSE              1677              304
2:       71 Upper Nile       01   Renk       01      003       FALSE              1104              184
3:       71 Upper Nile       01   Renk       01      004       FALSE               161               31
4:       71 Upper Nile       01   Renk       02      001       FALSE                87               16
5:       71 Upper Nile       01   Renk       02      003       FALSE               796              159
6:       71 Upper Nile       01   Renk       02      006       FALSE               419               82

mt12_est_rets_ind mt12_est_rets_hh    PAYAM      BOMA    GEOID9
1:              5087             1010 Chemmedi Bebineeth 710101001
2:              9446             1892 Chemmedi  Chemmedi 710101003
3:              1766              347 Chemmedi    Labior 710101004
4:              6854             1349    Geger   Gerbena 710102001
5:             26238             5147    Geger  Gospamni 710102003
6:              5967             1195    Geger   Wunthow 710102006

这个错误的核心是sample()函数接收的第一个参数无效——要么是空向量，要么不是可抽样的对象。结合你的代码和数据，问题主要出在这几个地方：

• County名称不匹配：idp.breakdown里的部分County（比如"Abyei"）在idp.frame中没有对应数据，导致筛选后返回空向量，无法抽样。
• %nin%运算符依赖问题：如果未加载data.table或dplyr，这个运算符未定义，会导致筛选未抽样数据时出错。
• 缺少空值判断：没有提前检查待抽样的数据集是否为空，直接执行sample()触发报错。

可以按以下步骤修复：

1. 检查并修复County匹配问题
   先找出idp.breakdown中在idp.frame里没有对应数据的County：

   missing_counties <- setdiff(idp.breakdown$County, idp.frame$County)
   print(missing_counties)
   

   针对这些County，要么补充idp.frame的数据，要么在函数中跳过这些行。

2. 替换%nin%为基础R写法
   用!GEOID9 %in% main.sample替代GEOID9 %nin% main.sample，避免依赖第三方包的运算符：

   # 修正后的预留抽样部分
   reserve.sample <- sample(
     idp.frame[COUNTY == idp.breakdown[x, County] & !GEOID9 %in% main.sample, GEOID9],
     n.reserve,
     prob = idp.frame[COUNTY == idp.breakdown[x, County] & !GEOID9 %in% main.sample, mt12_est_idps_hh],
     replace = F 
   )
   

3. 在函数中添加空值判断
   在执行抽样前先检查数据集是否为空，避免报错：

   idpSample <- function(x) {
     # 获取当前County的数据集
     county_data <- idp.frame[COUNTY == idp.breakdown[x, County]]
     if(nrow(county_data) == 0) {
       warning(paste("无对应数据的County:", idp.breakdown[x, County]))
       return(data.frame(GEOID9=character(), sample_type=character()))
     }
     
     # 主抽样
     main.sample <- sample(
       county_data$GEOID9,
       idp.breakdown[x, idp_cluster],
       prob = county_data$mt12_est_idps_hh,
       replace = T 
     )
     
     # 计算预留抽样数量
     n.reserve <- min(3, max(0, nrow(county_data) - length(main.sample)))
     reserve.sample <- c()
     if(n.reserve > 0) {
       remaining_data <- county_data[!GEOID9 %in% main.sample]
       if(nrow(remaining_data) >= n.reserve) {
         reserve.sample <- sample(
           remaining_data$GEOID9,
           n.reserve,
           prob = remaining_data$mt12_est_idps_hh,
           replace = F 
         )
       }
     }
     
     # 返回结果
     data.frame(
       GEOID9 = c(main.sample, reserve.sample),
       sample_type = c(rep("Main sample", length(main.sample)),
                      rep("Reserve", length(reserve.sample)))
     )
   }
   

4. 验证抽样参数
   确认idp.breakdown$idp_cluster都是正整数，避免传入0或负数导致抽样失败。

内容的提问来源于stack exchange，提问作者Emmanuel