[英]Split, reshape, bind stacked wide data using tidyverse in single pipe
我有一個單一的數據框,它為多個地區和村庄堆積了大量數據,其中有政黨記錄的投票。 每個地區都有不同的政黨:
df_in <- data.frame(
X1 = c(rep("District1", 3), rep("District2", 3)),
X2 = c(rep(c("", "Village1", "Village2"), 2)),
X3 = c("Party1", "30", "11", "Party1", "2", "59"),
X4 = c("Party2", "55", "42", "Party2", "66", "44"),
X5 = c("", "", "", "Party3", "32", "13"),
X6 = c("", "", "", "Party4", "99", "75")
)
我想最終為每個村/區的每一方記錄一個長的投票數據集:
df_out <- data.frame(
X1 = c(rep("District1", 4), rep("District2", 8)),
X2 = c("Village1", "Village1", "Village2", "Village2", "Village1", "Village1", "Village1", "Village1", "Village2", "Village2", "Village2", "Village2"),
X3 = c(
rep(c("Party1", "Party2"), 2),
rep(c("Party1", "Party2", "Party3", "Party4"), 2)
),
X4 = c(30, 55, 11, 42, 2, 66, 32, 99, 59, 44, 13, 75)
)
我想在單個管道中從輸入到輸出。 我一直在做類似以下設置的事情,但到目前為止沒有成功:
df_out <- df_in %>%
split(.$X1) %>%
map() %>%
gather() %>%
bind_rows()
這是正確的嗎?
我們也可以這樣做
library(dplyr)
library(tidyr)
library(hablar)
df_in %>%
# group by 'X1'
group_by(X1) %>%
# remove the first row
slice(-1) %>%
# ungroup
ungroup %>%
# rename the column names with 'Party'
rename_at(vars(X3:X6), ~ paste0("Party", 1:4)) %>%
# change the type of columns
retype %>%
# gather into long format
gather(X3, X4, Party1:Party4, na.rm = TRUE) %>%
# arrange if needed
arrange(X1, X2)
# A tibble: 12 x 4
# X1 X2 X3 X4
# <chr> <chr> <chr> <int>
# 1 District1 Village1 Party1 30
# 2 District1 Village1 Party2 55
# 3 District1 Village2 Party1 11
# 4 District1 Village2 Party2 42
# 5 District2 Village1 Party1 2
# 6 District2 Village1 Party2 66
# 7 District2 Village1 Party3 32
# 8 District2 Village1 Party4 99
# 9 District2 Village2 Party1 59
#10 District2 Village2 Party2 44
#11 District2 Village2 Party3 13
#12 District2 Village2 Party4 75
library(tidyverse)
df_in %>%
split(.$X1) %>%
map(. %>% gather(key,val,X3:X6) %>%
group_by(key) %>% mutate(key1=first(val)) %>% filter(row_number() %in% 2:n() & val!="") %>%
ungroup() %>% rename(X4=val, X3=key1) %>% select(X1,X2,X3,X4)) %>%
bind_rows()
更短,並具有適當的列名稱。
library(tidyr)
library(dplyr)
library(magrittr)
df_in %>%
mutate_all(as.character) %>% # Or set stringsAsFactors = FALSE
set_names(c("district", "village", paste0("Party", 1:4))) %>%
filter(nchar(village) > 0) %>%
gather(party, votes, -district, -village) %>%
mutate(votes = as.integer(votes) %>% replace_na(0)) %>%
arrange(district, village, party) %>%
filter(votes > 0)
## district village party votes
## 1 District1 Village1 Party1 30
## 2 District1 Village1 Party2 55
## 3 District1 Village2 Party1 11
## 4 District1 Village2 Party2 42
## 5 District2 Village1 Party1 2
## 6 District2 Village1 Party2 66
## 7 District2 Village1 Party3 32
## 8 District2 Village1 Party4 99
## 9 District2 Village2 Party1 59
## 10 District2 Village2 Party2 44
## 11 District2 Village2 Party3 13
## 12 District2 Village2 Party4 75
聲明:本站的技術帖子網頁,遵循CC BY-SA 4.0協議,如果您需要轉載,請注明本站網址或者原文地址。任何問題請咨詢:yoyou2525@163.com.