如何在R中旋转包含带有部分和子部分的列的数据框
Posted
技术标签:
【中文标题】如何在R中旋转包含带有部分和子部分的列的数据框【英文标题】:How to pivoting dataframe consisting column with section and sub section In R 【发布时间】:2020-07-23 12:06:08 【问题描述】:我有一个下面提到的数据框:
structure(
list(ID = c("P-1", " P-1", "P-1", "P-2", "P-3", "P-4", "P-5", "P-6", "P-7",
"P-8"),
Date = c("2020-03-16 12:11:33", "2020-03-16 13:16:04",
"2020-03-16 06:13:55", "2020-03-16 10:03:43",
"2020-03-16 12:37:09", "2020-03-16 06:40:24",
"2020-03-16 09:46:45", "2020-03-16 12:07:44",
"2020-03-16 14:09:51", "2020-03-16 09:19:23"),
Status = c("SA", "SA", "SA", "RE", "RE", "RE", "RE", "XA", "XA", "XA"),
Flag = c("L", "L", "L", NA, "K", "J", NA, NA, "H", "G"),
Value = c(5929.81, 5929.81, 5929.81, NA, 6969.33, 740.08, NA, NA, 1524.8,
NA),
Flag2 = c("CL", "CL", "CL", NA, "RY", "", NA, NA, "", NA),
Flag3 = c(NA, NA, NA, NA, "RI", "PO", NA, "SS", "DDP", NA)),
.Names=c("ID", "Date", "Status", "Flag", "Value", "Flag2", "Flag3"),
row.names=c(NA, 10L), class="data.frame")
我正在使用下面提到的代码:
df %>% mutate(L = ifelse(Flag == "L",1,0),
K = ifelse(Flag == "K",1,0),
# etc for Flag) %>%
mutate(sub_status = NA) %>%
mutate(sub_status = ifelse(!is.na(Flag2) & Flag3 == 0, "a", sub_status),
sub_status = ifelse(is.na(Flag2) & Flag3 != 0, "b", sub_status),
# etc for sub-status) %>%
mutate(value_class = ifelse(0 <= Value & Value <= 15000, "0-15000",
"15000-50000")) %>%
group_by(Date, status, sub_status, value_class) %>%
summarise(L = sum(L),
K = sum(K),
# etc
count = n())
它为我提供了以下输出:
Date Status sub_status value_class G H I J K L NA Count
2020-03-20 SA a 0-15000 0 0 0 0 1 1 0 2
2020-03-20 SA b 0-15000 0 0 0 0 1 0 0 1
................
................
我想使用DF
获得以下输出,其中Status
列具有不同的 3 个值,Flag2
具有值或 [null] 或 NA,最后Flag3
列具有不同的 7 个值[null] 或 NA。对于一个不同的ID
,我们有多个Flag3
列条目。
我需要创建以下数据框,方法是创建一个基于 Value
的 3 个组,例如 0-15000、15000-50000。
Flag2
的值不是 0 或 [null]/NA,但 Flag3
的值是 0 或 [null]/NA,那么它将是 a
。
如果对于不同的 ID,Flag3
的值不是 0 或 [null]/NA,但 Flag2
的值是 0 或 [null]/NA,那么它将是 b
如果对于不同的 ID,Flag2
和 Flag3
的值都不是 0 或 [Null]/NA,那么它将是 c
如果 Flag2
和 Flag3
的值均为 0 或 [Null]/NA 对于不同的 ID,则为 d
我想用percent
和Total
列将上面提到的datafrmae安排在下面的结构中。
我提到了像2/5
这样的百分比,以表明状态将除以总数,而sub_status
将除以各自的Status
。
16/03/2020 0 - 15000 15000 - 50000
Status count percent L K J H G [Null] count percent L K J H G [Null] Total
SA 1 1/8 (12.50%) 1 0 0 0 0 0 0 - 0 0 0 0 0 0 1
a 1 1/1(100.00%) 1 0 0 0 0 0 0 - 0 0 0 0 0 0 1
b 0 - 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0
c 0 - 1 0 0 0 0 0 0 - 0 0 0 0 0 0 0
d 0 - 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0
RE 4 50.00% 0 1 1 0 0 2 0 - 0 0 0 0 0 0 4
a 0 - 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0
b 1 25.00% 0 0 1 0 0 1 0 - 0 0 0 0 0 0 1
c 1 25.00% 0 1 0 0 0 1 0 - 0 0 0 0 0 0 1
d 2 50.00% 0 0 0 0 0 2 0 - 0 0 0 0 0 0 2
XA 3 37.50% 0 0 0 1 1 1 0 - 0 0 0 0 0 0 3
a 0 - 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0
b 2 66.67% 0 0 0 1 0 1 0 - 0 0 0 0 0 0 2
c 0 - 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0
d 1 33.33% 0 0 0 0 1 0 0 - 0 0 0 0 0 0 1
Total 8 100.00% 1 1 0 0 1 3 0 - 0 0 0 0 0 0 8
我已经根据最新日期(2020 年 3 月 16 日)提到了所需的输出,如果数据帧没有按照startdate
的最新日期,则在输出数据帧中保留所有值 0。百分比列仅供参考,会有计算出的百分比值。
另外,我想保持结构不变。例如,如果某一天没有任何参数,则输出结构将与 0 值相同。
例如,假设日期17/03/2020
没有任何状态为SA
或子状态c
的行,其占位符将在输出中,值为0
。
【问题讨论】:
@akrun:我保留的百分比列就像2/5
只是为了表示目的。只有带百分号的小数点后两位的百分比值。
@akrun: 请建议是否可以通过 R:(
您的数据输入是 10 行,但预期 iis 更多。是否基于输入示例的预期
@akrun:很抱歉,输出仅用于视觉表示。我需要了解获得此类输出的方法。
您可以从您喜欢的数据集的dput
开始吗?它是第三个代码块。前面的代码看起来并不相关,因为您似乎对输出感到满意。
【参考方案1】:
希望这足以让您入门,更进一步,我需要一个看起来像是来自 R 的预期输出,并进一步解释如何计算变量。
library(tidyverse)
df <- structure(
list(ID = c("P-1", " P-1", "P-1", "P-2", "P-3", "P-4", "P-5", "P-6", "P-7",
"P-8"),
Date = c("2020-03-16 12:11:33", "2020-03-16 13:16:04",
"2020-03-16 06:13:55", "2020-03-16 10:03:43",
"2020-03-16 12:37:09", "2020-03-16 06:40:24",
"2020-03-16 09:46:45", "2020-03-16 12:07:44",
"2020-03-16 14:09:51", "2020-03-16 09:19:23"),
Status = c("SA", "SA", "SA", "RE", "RE", "RE", "RE", "XA", "XA", "XA"),
Flag = c("L", "L", "L", NA, "K", "J", NA, NA, "H", "G"),
Value = c(5929.81, 5929.81, 5929.81, NA, 6969.33, 740.08, NA, NA, 1524.8,
NA),
Flag2 = c("CL", "CL", "CL", NA, "RY", "", NA, NA, "", NA),
Flag3 = c(NA, NA, NA, NA, "RI", "PO", NA, "SS", "DDP", NA)),
.Names=c("ID", "Date", "Status", "Flag", "Value", "Flag2", "Flag3"),
row.names=c(NA, 10L), class="data.frame")
df2 <- df %>%
mutate(
# add variables
Value = ifelse(0 <= Value & Value <= 15000, "0-15000", "15000-50000"),
substatus = case_when(
!is.na(Flag2) & is.na(Flag3) ~ "a",
!is.na(Flag3) & is.na(Flag2) ~ "b",
!is.na(Flag3) & !is.na(Flag2) ~ "c",
TRUE ~ "d"),
# make Date an actual date rather than a timestamp
Date = as.Date(Date),
# remove obsolete columns
Flag2 = NULL,
Flag3 = NULL,
ID = NULL,
# renames NAs into the name of the desired column
Flag = ifelse(is.na(Flag), "[Null]", Flag),
# create column of 1 for pivot
temp = 1,
# and row id
id = row_number()
) %>%
# create new columns L K etc, this also drops the Flag col
pivot_wider(names_from = "Flag", values_from = "temp", values_fill = list(temp=0)) %>%
# move `[Null]` column to the end
select(everything(), -`[Null]`, `[Null]`) %>%
mutate(
id = NULL,
count = 1,
Total = rowSums(select(., L:`[Null]`)))
df2
#> # A tibble: 10 x 12
#> Date Status Value substatus L K J H G `[Null]`
#> <date> <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 2020-03-16 SA 0-15~ a 1 0 0 0 0 0
#> 2 2020-03-16 SA 0-15~ a 1 0 0 0 0 0
#> 3 2020-03-16 SA 0-15~ a 1 0 0 0 0 0
#> 4 2020-03-16 RE <NA> d 0 0 0 0 0 1
#> 5 2020-03-16 RE 0-15~ c 0 1 0 0 0 0
#> 6 2020-03-16 RE 0-15~ c 0 0 1 0 0 0
#> 7 2020-03-16 RE <NA> d 0 0 0 0 0 1
#> 8 2020-03-16 XA <NA> b 0 0 0 0 0 1
#> 9 2020-03-16 XA 0-15~ c 0 0 0 1 0 0
#> 10 2020-03-16 XA <NA> d 0 0 0 0 1 0
#> # ... with 2 more variables: count <dbl>, Total <dbl>
# As you didn't tell what to do with NA values so I left them as NA
bind_rows(
df2 %>%
# add missing combinations of abcd
complete(nesting(Date, Status, Value), substatus) %>%
group_by(Date, Value, Status, substatus) %>%
summarize_all(~sum(., na.rm=TRUE)) %>%
group_by(Status, Value) %>%
mutate(percent = paste(round(100 * Total / sum(Total), 2), "%")) %>%
ungroup(),
df2 %>%
mutate(substatus = Status, Status = paste0(Status, "_")) %>%
group_by(Date, Value, Status, substatus) %>%
mutate(count = n()) %>%
group_by(count, add = TRUE) %>%
summarize_all(~sum(., na.rm=TRUE)) %>%
group_by(Value) %>%
mutate(percent = paste(round(100 * Total / sum(Total), 2), "%"))
) %>%
arrange(Date, Value, desc(Status)) %>%
mutate(Status = NULL) %>%
rename(Status = substatus) %>%
print(n=Inf)
#> # A tibble: 25 x 12
#> Date Value Status L K J H G `[Null]` count Total
#> <date> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 2020-03-16 0-15~ XA 0 0 0 1 0 0 1 1
#> 2 2020-03-16 0-15~ a 0 0 0 0 0 0 0 0
#> 3 2020-03-16 0-15~ b 0 0 0 0 0 0 0 0
#> 4 2020-03-16 0-15~ c 0 0 0 1 0 0 1 1
#> 5 2020-03-16 0-15~ d 0 0 0 0 0 0 0 0
#> 6 2020-03-16 0-15~ SA 3 0 0 0 0 0 3 3
#> 7 2020-03-16 0-15~ a 3 0 0 0 0 0 3 3
#> 8 2020-03-16 0-15~ b 0 0 0 0 0 0 0 0
#> 9 2020-03-16 0-15~ c 0 0 0 0 0 0 0 0
#> 10 2020-03-16 0-15~ d 0 0 0 0 0 0 0 0
#> 11 2020-03-16 0-15~ RE 0 1 1 0 0 0 2 2
#> 12 2020-03-16 0-15~ a 0 0 0 0 0 0 0 0
#> 13 2020-03-16 0-15~ b 0 0 0 0 0 0 0 0
#> 14 2020-03-16 0-15~ c 0 1 1 0 0 0 2 2
#> 15 2020-03-16 0-15~ d 0 0 0 0 0 0 0 0
#> 16 2020-03-16 <NA> XA 0 0 0 0 1 1 2 2
#> 17 2020-03-16 <NA> a 0 0 0 0 0 0 0 0
#> 18 2020-03-16 <NA> b 0 0 0 0 0 1 1 1
#> 19 2020-03-16 <NA> c 0 0 0 0 0 0 0 0
#> 20 2020-03-16 <NA> d 0 0 0 0 1 0 1 1
#> 21 2020-03-16 <NA> RE 0 0 0 0 0 2 2 2
#> 22 2020-03-16 <NA> a 0 0 0 0 0 0 0 0
#> 23 2020-03-16 <NA> b 0 0 0 0 0 0 0 0
#> 24 2020-03-16 <NA> c 0 0 0 0 0 0 0 0
#> 25 2020-03-16 <NA> d 0 0 0 0 0 2 2 2
#> # ... with 1 more variable: percent <chr>
【讨论】:
非常感谢,您能否帮助根据价值对框架进行分类(即0-15
、15-50
和50+
)。另外,我怎样才能得到所需的百分比列,以上是关于如何在R中旋转包含带有部分和子部分的列的数据框的主要内容,如果未能解决你的问题,请参考以下文章
如何使用 oracle sql 命令查找 clob 数据列的特定子部分(实际存储字符)?
如何在整个 Pandas 数据框中搜索字符串并获取包含它的列的名称?