将函数应用于满足data.table中列子集条件的子集元素
[英]Apply function to subset elements that satisfy a criteria across subset of columns in data.table
问题描述
Consider the data.table, exampleDT , 
考虑data.table， exampleDT ，
set.seed(7)
exampleDT = data.table(colA = rnorm(10,15,5),
colB = runif(10,100,150),
targetA = rnorm(10,12,2),
targetB = rnorm(10,8,4))
If I want to calculate the mean of all elements in column targetA , for example, that are below some threshold -- say, 10 -- I can do the following: 例如，如果我要计算targetA列中所有元素的均值，这些均值低于某个阈值（例如10），则可以执行以下操作：
examp_threshold = 10
exampleDT[targetA<examp_threshold,mean(targetA)]
# [1] 9.224007566814299
And if I want to calculate the mean of all elements in columns targetA and targetB , for example, I can do the following: 而且，例如，如果我要计算targetA和targetB列中所有元素的均值，则可以执行以下操作：
target_cols = names(exampleDT)[which(names(exampleDT) %like% "target")]
exampleDT[,lapply(.SD,mean),.SDcols=target_cols]
# targetA targetB
# 1: 12.60101574551183 7.585007905896557
But I don't know how to combine the two; 但是我不知道如何将两者结合起来。 that is, to calculate the mean of all elements in all columns containing a specified string ("target", in this case) that are below some specified threshold (10, here). 也就是说，计算包含指定字符串（在此情况下为“目标”）的所有列中的所有元素的均值低于某个指定阈值（此处为10）的平均值。 This was my first guess, but it was unsuccessful: 这是我的第一个猜测，但未成功：
exampleDT[.SD<examp_threshold,lapply(.SD,mean),.SDcols=target_cols]
#Empty data.table (0 rows) of 2 cols: targetA,targetB
1 个解决方案
解决方案1
2 已采纳 2017-07-20 13:59:14
You need to subset in the j expression, like so: 您需要在j表达式中设置子集，如下所示：
exampleDT[, lapply(.SD, function(x) mean(x[x<examp_threshold])),.SDcols=target_cols]
# targetA targetB
#1: 9.224008 6.66624
声明:本站的技术帖子网页，遵循CC BY-SA 4.0协议，如果您需要转载，请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.