[英]How to create new column and calculate median in data.table when using a subset of data
Hi I have a data like this: 嗨,我有这样的数据:
date type data
198101 1 0.1
198101 1 0.3
198101 2 0.5
198102 1 1.2
198102 1 0.9
198102 2 0.7
198102 2 0.3
I would like to create a new column to show the median each month according to criteria when type == 1. 我想创建一个新列,以在类型== 1时根据条件显示每月的中位数。
The result I would like to be is like this 我想要的结果是这样的
date type data P50
198101 1 0.1 0.2
198101 1 0.3 0.2
198101 2 0.5 0.2
198102 1 1.2 1.05
198102 1 0.9 1.05
198102 2 0.7 1.05
198102 2 0.3 1.05
currently I do it this way, lets call the above data.table as dt 目前我是用这种方式,让我们将上面的data.table称为dt
dt.median = dt[type == 1]
dt.median = dt.median[, .(P50 = median(data)), by=.(date)]
Then merge it back into the original dt 然后将其合并回原始dt
dt = dt[dt.median, nomatch = 0, by=.(date)]
Is there a quicker way to do this using .SD or .SDcol? 是否有使用.SD或.SDcol的更快方法? I want to practice using .SD but just cannot figure it out with maybe one line of code?
我想练习使用.SD,但可能无法用一行代码来解决?
What I could think of is currently 我现在能想到的
dt[, P50 := * .SD[type == 1] ... * , by =.(date)]
but then I dont know what the syntax to put in to calculate median * .SD[type == 1] ... *, 但后来我不知道要使用什么语法来计算中位数* .SD [type == 1] ... *,
Help will be much appreciated! 帮助将不胜感激!
Just index the data values within groups using a logical vector and assign with the data.table special assignment operator, :=
只需使用逻辑向量为组内的数据值编制索引,并使用data.table特殊赋值运算符
:=
> dt[ , P50 := median(data[type==1]), by=.(date)]
> dt
date type data P50
1: 198101 1 0.1 0.20
2: 198101 1 0.3 0.20
3: 198101 2 0.5 0.20
4: 198102 1 1.2 1.05
5: 198102 1 0.9 1.05
6: 198102 2 0.7 1.05
7: 198102 2 0.3 1.05
From base R 从基数R
v=dt$data
v[dt$type!=1]=NA
ave(v,dt$date,FUN=function(x) median(x,na.rm=T))
[1] 0.20 0.20 0.20 1.05 1.05 1.05 1.05
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.