dcast将所有变量重命名为以数字开头 [英] dcast renaming all variables to start with a number
问题描述
因此，我得到的数据如下：
So I've got data that looks like this:
id year principal interest
1: 011000600 2013 0.00 0.00
2: 011000600 2014 544.03 0.00
3: 011000700 2013 0.00 0.00
4: 011000700 2014 0.01 0.00
5: 011000800 2013 363.44 12.79
6: 011000800 2014 2005.98 0.00
7: 011000900 2013 0.00 0.00
8: 011000900 2014 0.00 0.00
9: 011001000 2013 0.00 0.00
10: 011001000 2014 0.00 0.00
11: 011001100 2013 0.00 0.00
12: 011001100 2014 1723.24 0.00
13: 011001560 2013 0.00 0.00
14: 011001560 2014 0.00 0.00
15: 011001650 2013 0.00 0.00
16: 011001650 2014 0.00 0.00
（基本上是一堆变量的纵向示例）
(basically a longitudinal sample of a bunch of variables)
数据在很大的一面，所以我使用 data.table 。我重塑它以获得每个 id 独特的行：
The data is on the large side so I'm using data.table for everything. I reshape it to get each id unique by row:
datam<-melt(data,id=c("id","year"))
data1<-dcast.data.table(datam,id~...)
这会产生：
id 2013_principal 2013_interest 2014_principal 2014_interest
1: 011000600 0.00 0.00 544.03 0
2: 011000700 0.00 0.00 0.01 0
3: 011000800 363.44 12.79 2005.98 0
4: 011000900 0.00 0.00 0.00 0
5: 011001000 0.00 0.00 0.00 0
6: 011001100 0.00 0.00 1723.24 0
我想要的数据的任何建议如何处理这个？我更愿意：
Any suggestions for how to deal with this? I'd much rather have:
id principal_2013 interest_2013 principal_2014 interest_2014
1: 011000600 0.00 0.00 544.03 0
2: 011000700 0.00 0.00 0.01 0
3: 011000800 363.44 12.79 2005.98 0
4: 011000900 0.00 0.00 0.00 0
5: 011001000 0.00 0.00 0.00 0
6: 011001100 0.00 0.00 1723.24 0
（将年更改为后缀）
我试图更多显示投放时，例如
(switching the year to be a suffix) I've tried being more explicit when casting, e.g.
data2<-dcast.data.table(datam,id~year+...)
data3<-dcast.data.table(datam,id~...+year)
无效：
data2
id 2013_principal 2013_interest 2014_principal 2014_interest
1: 011000600 0.00 0.00 544.03 0
2: 011000700 0.00 0.00 0.01 0
3: 011000800 363.44 12.79 2005.98 0
4: 011000900 0.00 0.00 0.00 0
5: 011001000 0.00 0.00 0.00 0
6: 011001100 0.00 0.00 1723.24 0
data3
id 2013_principal 2013_interest 2014_principal 2014_interest
1: 011000600 0.00 0.00 544.03 0
2: 011000700 0.00 0.00 0.01 0
3: 011000800 363.44 12.79 2005.98 0
4: 011000900 0.00 0.00 0.00 0
5: 011001000 0.00 0.00 0.00 0
6: 011001100 0.00 0.00 1723.24 0
看来dcast的命名约定默认为这种风格，假设我想象这种类型的重塑是无所不在的。
Seems pretty silly for the naming convention of dcast to default to this style, given that I imagine this type of reshaping is ubiquitous.
我也尝试过补丁的事后给一些其他的帖子，我发现（例如此处），但它运行速度非常慢（有〜 400个变量以在完整数据集中重命名）
I've also tried patching things up ex-post given some other posts I've found (e.g. here), but it runs unfathomably slow (there are ~400 variables to rename in the full data set)
names(data)<-ifelse(substr(names(data),1,2) %in% c("19","20"),
paste(substr(names(data),6,nchar(data)),
substr(names(data),1,4),sep="_") ,
names(copy))
（我试图找到以年 - 19xx或20xx开始的所有变量 - 并尝试交换开始和结束）
(I'm trying to find all the variables starting with years--19xx or 20xx--and trying to swap the beginning and end)
推荐答案
更好的是，在 dcast wiki / Installationrel =nofollow> v1.9.5 + ，我们可以同时投放多个列。
Even better, with the new developments to dcast in data.table from v1.9.5+, we can cast multiple columns simultaneously..
require(data.table) # v1.9.5+
dcast(dt, id ~ year, value.var = c("principal", "interest"))
# id principal_2013 principal_2014 interest_2013 interest_2014
# 1: 11000600 0.00 544.03 0.00 0
# 2: 11000700 0.00 0.01 0.00 0
# 3: 11000800 363.44 2005.98 12.79 0
# 4: 11000900 0.00 0.00 0.00 0
# 5: 11001000 0.00 0.00 0.00 0
# 6: 11001100 0.00 1723.24 0.00 0
# 7: 11001560 0.00 0.00 0.00 0
# 8: 11001650 0.00 0.00 0.00 0
不必再有融化 > cast ，因此效率很高。
No more having to melt unnecessarily before to cast, therefore quite efficient.
这篇关于dcast将所有变量重命名为以数字开头的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！
