使用以下data.table：
dt_cpu_[2000:2010,]
tim source type_instance val
1: 2016-08-22 05:16:12.336 DBNode3 system 29471
2: 2016-08-22 05:35:46.788 DBNode1 steal 0
3: 2016-08-22 05:16:12.336 DBNode3 steal 0
4: 2016-08-22 05:35:46.788 DBNode1 nice 617733
5: 2016-08-22 05:16:12.336 DBNode3 wait 452
6: 2016-08-22 05:16:12.336 DBNode3 idle 401793528
7: 2016-08-22 05:16:12.336 DBNode3 steal 0
8: 2016-08-22 05:16:02.336 DBNode3 softirq 0
9: 2016-08-22 05:16:02.336 DBNode3 steal 0
10: 2016-08-22 05:36:28.926 DNSFront wait 211213
11: 2016-08-22 05:36:28.926 DNSFront nice 661508
使用以下类型：
str(dt_cpu_)
Classes ‘data.table’ and 'data.frame': 10576673 obs. of 4 variables:
$ tim : POSIXct, format: "2016-08-22 05:06:48.901"
$ source : chr "DNSBack" "DBNode3" "DNSBack" "DBNode2" ...
$ type_instance: chr "softirq" "user" "system" "user" ...
$ val : num 0 99732 829061 171631 15111 ...
我使用以下代码来转换data.table：
dt_cpu_cast <- dcast.data.table(dt_cpu_[2000:2010,], tim + source ~ type_instance, drop = TRUE, value.var="val")
我希望得到如下输出：
tim source system steal nice wait idle softirq
1: 2016-08-22 05:16:12.336 DBNode3 29471 0 NA 452 401793528 NA
1: 2016-08-22 05:35:46.788 DBNode1 NA 0 617733 NA NA NA
但我得到以下内容：
tim source idle nice softirq steal system wait
1: 2016-08-22 05:16:02.336 DBNode3 0 0 1 1 0 0
2: 2016-08-22 05:16:12.336 DBNode3 1 0 0 2 1 1
3: 2016-08-22 05:35:46.788 DBNode1 0 1 0 1 0 0
4: 2016-08-22 05:36:28.926 DNSFront 0 1 0 0 0 1
我不明白我做错了什么。你能给我一些指导吗？