Question

我正在尝试在多个条件下对一个大数据框（北半球的气旋轨道）进行子集化：下面的数据

centro <- read.table("https://forms.naturwissenschaften.ch/imilast/_ERAinterim_1.5_1979_MTEX/ERAinterim_1.5_NH_M02_19790101_20121231_MTEX.txt?_ga=2.18919096.1825595846.1546710263-1112023567.1546710263", sep="", fill = T, nrows = 500,
                 header = F, skip = 2) # read here only the first 500 rows

centro <- na.omit(centro)

colnames(centro) <- c("Code","CycloneNo","StepNo","DateI10","Year","Month","Day","Time","LongE","LatN","Intensity1","Intensity2","Intensity3")

当列StepNo == 1时，我只想对空间框中（如-4和40 E经度和32-45 N lat）形成的旋风器（基于唯一列CycloneNo）进行子集化。通常，这很容易做到：

centro_subs <- centro[centro$LongE>=-4 & centro$LongE <= 40 & centro$LatN>= 32 & centro$LatN <= 45,]

但是，我只想保留在此框中形成的旋风分离器（当StepNo == 1时），而其余轨道也都保留在此框之外。

我试图这样做：

df_s <- centro[1,]
df_s[1,] <- NA # create an empty dataframe to be filled in the iteration


for (i in 1:length(unique(centro$CycloneNo))){
print(i)
a <- centro[centro$LongE[centro$StepNo==1]>= -4 & 
centro$LongE[centro$StepNo==1] <= 40 & 
centro$LatN[centro$StepNo==1]>= 32 & centro$LatN 
<=45[centro$StepNo==1],]
df_s <- rbind(a, df_s)
}

但是，最终结果是在一个充满NA的空数据框中。我知道这很难在这里描述。我觉得自己有点接近，但是我现在也很疲惫，试图找到新的方法。

Answer 1

我不认为您想要循环播放。我敢肯定这不是最优雅的方法，但是我认为它是可行的。

step1s <- subset(centro, StepNo == 1) # only take step 1 of all cyclones
keeps <- step1s$CycloneNo[step1s$LongE>=-4 & step1s$LongE <= 40 & step1s$LatN>= 32 & step1s$LatN <= 45] # find cyclone numbers for cyclones meeting the condition
centro_sub <- centro[centro$CycloneNo %in% keeps, ] # keep all steps of cyclones meeting the conditions

Answer 2

约瑟夫提供了一个很好的答案。或者，可以在data.table中使用它，这可能会以某种速度为代价提供更高的可读性。

centro <- data.table(centro)
centro[CycloneNo %in% CycloneNo[StepNo == 1 & 
                                  LongE %between% c(-4,40) & 
                                  LatN %between% c(32,45)]]

基于R中的几种条件的子集数据帧

2 个答案: