我有以下数据框:
Group User Manager Client Other data1 Other data2
AG TRUE TRUE FALSE a b
AG TRUE TRUE FALSE c d
JU FALSE FALSE TRUE d e
JU TRUE TRUE FALSE f g
我想使用像这样的结果:
USER MANAGER CLIENT COUNT-AG COUNT-JU
TRUE TRUE FALSE 2 1
FALSE FALSE TRUE 0 1
我只能使用以下内容获得一组或另一组的结果:
f = filter(dataframe, Group == “AG”)
result <- group_by(f,User,Manager) %>%
summarize(count=n())
但我无法找到一种方法来获得同一结果中两组的计数。 有什么想法吗?
答案 0 :(得分:0)
我们可以使用spread
中的tidyr
来重塑“长期&#39;广泛的&#39;包括&#39; Group&#39;之后的格式也在group_by
。
library(dplyr)
library(tidyr)
df1 %>%
group_by(User, Manager, Client, Group) %>%
summarise(Count = n()) %>%
spread(Group, Count, fill=0)%>%
rename(COUNT_AG=AG, COUNT_JU=JU)
# User Manager Client COUNT_AG COUNT_JU
# (lgl) (lgl) (lgl) (dbl) (dbl)
#1 FALSE FALSE TRUE 0 1
#2 TRUE TRUE FALSE 2 1
如果我们使用data.table
,请转换&#39; data.frame&#39;到&#39; data.table&#39; (setDT(df1)
,并将dcast
与fun.aggregate
一起用作length
。
library(data.table)
dcast(setDT(df1), User+Manager+Client~Group, length)
df1 <- structure(list(Group = c("AG", "AG", "JU", "JU"),
User = c(TRUE,
TRUE, FALSE, TRUE), Manager = c(TRUE, TRUE, FALSE,
TRUE), Client = c(FALSE,
FALSE, TRUE, FALSE), Other_data1 = c("a", "c", "d",
"f"), Other_data2 = c("b",
"d", "e", "g")), .Names = c("Group", "User", "Manager",
"Client",
"Other_data1", "Other_data2"), class = "data.frame",
row.names = c(NA, -4L))
答案 1 :(得分:0)
正如我在评论中提到的,你必须使用data.table。这是在第一行完成的(+计数由.N完成)。第二行符合您的要求:
yyyy-mm-dd