如何防止group_split()自动排序并保留原始顺序

Question

1

假设我下载了一个名为econ_dmg.csv的数据集

fileUrl <- "https://raw.githubusercontent.com/MichaelSodeke/DataSets/main/econ_dmg.csv"
download.file(fileUrl, destfile="econ_dmg.csv", method = "curl", mode="wb")
df <- read_csv("econ_dmg.csv") 

当前顺序为降序(按max.net.loss)，我希望在使用group_split后保留该顺序

> df
# A tibble: 91 x 3
   state evtype            max.net.loss
   <fct> <fct>                    <dbl>
 1 AM    WATERSPOUT                5000
 2 IL    FLASH FLOOD               5000
 3 NC    THUNDERSTORM WIND         5000
 4 PR    LANDSLIDE                 4800
 5 TN    TORNADO                   4410
 6 OK    THUNDERSTORM WIND         3500
 7 CT    THUNDERSTORM WIND         3200
 8 IN    FLOOD                     3000
 9 KY    FLASH FLOOD               3000
10 CA    LANDSLIDE                 1584
# ... with 81 more rows

2

然后按如下方式对元素进行分组：

df2 <- df %>% group_by(state, max.net.loss)

group_by保留顺序：

> df2
# A tibble: 91 x 3
# Groups:   state, max.net.loss [72]
   state evtype            max.net.loss
   <fct> <fct>                    <dbl>
 1 AM    WATERSPOUT                5000
 2 IL    FLASH FLOOD               5000
 3 NC    THUNDERSTORM WIND         5000
 4 PR    LANDSLIDE                 4800
 5 TN    TORNADO                   4410
 6 OK    THUNDERSTORM WIND         3500
 7 CT    THUNDERSTORM WIND         3200
 8 IN    FLOOD                     3000
 9 KY    FLASH FLOOD               3000
10 CA    LANDSLIDE                 1584
# ... with 81 more rows

3:问题

接下来，我使用以下命令将数据帧转换为列表：

df3 <- group_split(df2)

但是，group_split不考虑排序，因此按state变量而不是max.net.loss排序

> df3
<list_of<
  tbl_df<
    state       : factor<00fd4>
    evtype      : factor<70f44>
    max.net.loss: double
  >
>[72]>
[[1]]
# A tibble: 1 x 3
  state evtype        max.net.loss
  <fct> <fct>                <dbl>
1 AK    COASTAL FLOOD         1000

[[2]]
# A tibble: 1 x 3
  state evtype  max.net.loss
  <fct> <fct>          <dbl>
1 AL    TORNADO          960

[[3]]
# A tibble: 1 x 3
  state evtype     max.net.loss
  <fct> <fct>             <dbl>
1 AM    WATERSPOUT         5000

[[4]]
# A tibble: 1 x 3
  state evtype                   max.net.loss
  <fct> <fct>                           <dbl>
1 AN    MARINE THUNDERSTORM WIND          100

[[5]]
# A tibble: 1 x 3
  state evtype  max.net.loss
  <fct> <fct>          <dbl>
1 AR    TORNADO          950

[[6]]
# A tibble: 2 x 3
  state evtype         max.net.loss
  <fct> <fct>                 <dbl>
1 AS    HURRICANE               500
2 AS    TROPICAL STORM          500

[[7]]
# A tibble: 1 x 3
  state evtype max.net.loss
  <fct> <fct>         <dbl>
1 AZ    HAIL            900

[[8]]
# A tibble: 1 x 3
  state evtype    max.net.loss
  <fct> <fct>            <dbl>
1 CA    LANDSLIDE         1584

[[9]]
# A tibble: 1 x 3
  state evtype    max.net.loss
  <fct> <fct>            <dbl>
1 CO    HIGH WIND          979

[[10]]
# A tibble: 1 x 3
  state evtype            max.net.loss
  <fct> <fct>                    <dbl>
1 CT    THUNDERSTORM WIND         3200

[[11]]
# A tibble: 1 x 3
  state evtype    max.net.loss
  <fct> <fct>            <dbl>
1 DC    TSTM WIND          600

[[12]]
# A tibble: 1 x 3
  state evtype    max.net.loss
  <fct> <fct>            <dbl>
1 DE    LIGHTNING          800

[[13]]
# A tibble: 1 x 3
  state evtype max.net.loss
  <fct> <fct>         <dbl>
1 FL    FLOOD           935

[[14]]
# A tibble: 1 x 3
  state evtype    max.net.loss
  <fct> <fct>            <dbl>
1 GA    ICE STORM         952.

[[15]]
# A tibble: 1 x 3
  state evtype           max.net.loss
  <fct> <fct>                   <dbl>
1 GM    MARINE TSTM WIND          100

[[16]]
# A tibble: 1 x 3
  state evtype           max.net.loss
  <fct> <fct>                   <dbl>
1 GU    STORM SURGE/TIDE          900

[[17]]
# A tibble: 1 x 3
  state evtype   max.net.loss
  <fct> <fct>           <dbl>
1 HI    WILDFIRE          950

[[18]]
# A tibble: 1 x 3
  state evtype    max.net.loss
  <fct> <fct>            <dbl>
1 IA    ICE STORM          989

[[19]]
# A tibble: 1 x 3
  state evtype      max.net.loss
  <fct> <fct>              <dbl>
1 ID    FLASH FLOOD          800

[[20]]
# A tibble: 1 x 3
  state evtype      max.net.loss
  <fct> <fct>              <dbl>
1 IL    FLASH FLOOD         5000

[[21]]
# A tibble: 1 x 3
  state evtype max.net.loss
  <fct> <fct>         <dbl>
1 IN    FLOOD          3000

[[22]]
# A tibble: 1 x 3
  state evtype   max.net.loss
  <fct> <fct>           <dbl>
1 KS    WILDFIRE         1000

[[23]]
# A tibble: 1 x 3
  state evtype      max.net.loss
  <fct> <fct>              <dbl>
1 KY    FLASH FLOOD         3000

[[24]]
# A tibble: 1 x 3
  state evtype    max.net.loss
  <fct> <fct>            <dbl>
1 LA    ICE STORM          900

[[25]]
# A tibble: 3 x 3
  state evtype                   max.net.loss
  <fct> <fct>                           <dbl>
1 LC    MARINE HAIL                         0
2 LC    MARINE THUNDERSTORM WIND            0
3 LC    MARINE TSTM WIND                    0

[[26]]
# A tibble: 1 x 3
  state evtype                   max.net.loss
  <fct> <fct>                           <dbl>
1 LE    MARINE THUNDERSTORM WIND           25

[[27]]
# A tibble: 4 x 3
  state evtype                   max.net.loss
  <fct> <fct>                           <dbl>
1 LH    MARINE HAIL                         0
2 LH    MARINE THUNDERSTORM WIND            0
3 LH    MARINE TSTM WIND                    0
4 LH    WATERSPOUT                          0

[[28]]
# A tibble: 1 x 3
  state evtype           max.net.loss
  <fct> <fct>                   <dbl>
1 LM    MARINE TSTM WIND          500

[[29]]
# A tibble: 1 x 3
  state evtype           max.net.loss
  <fct> <fct>                   <dbl>
1 LO    MARINE TSTM WIND           25

[[30]]
# A tibble: 1 x 3
  state evtype           max.net.loss
  <fct> <fct>                   <dbl>
1 LS    MARINE TSTM WIND          400

[[31]]
# A tibble: 1 x 3
  state evtype    max.net.loss
  <fct> <fct>            <dbl>
1 MA    HIGH WIND          900

[[32]]
# A tibble: 2 x 3
  state evtype         max.net.loss
  <fct> <fct>                 <dbl>
1 MD    FLASH FLOOD             900
2 MD    TROPICAL STORM          900

[[33]]
# A tibble: 1 x 3
  state evtype max.net.loss
  <fct> <fct>         <dbl>
1 ME    FLOOD           954

[[34]]
# A tibble: 1 x 3
  state evtype    max.net.loss
  <fct> <fct>            <dbl>
1 MH    HIGH SURF            5

[[35]]
# A tibble: 1 x 3
  state evtype      max.net.loss
  <fct> <fct>              <dbl>
1 MI    FLASH FLOOD          970

[[36]]
# A tibble: 1 x 3
  state evtype max.net.loss
  <fct> <fct>         <dbl>
1 MN    FLOOD           996

[[37]]
# A tibble: 1 x 3
  state evtype max.net.loss
  <fct> <fct>         <dbl>
1 MO    FLOOD           940

[[38]]
# A tibble: 1 x 3
  state evtype max.net.loss
  <fct> <fct>         <dbl>
1 MS    FLOOD           955

[[39]]
# A tibble: 1 x 3
  state evtype max.net.loss
  <fct> <fct>         <dbl>
1 MT    HAIL            900

[[40]]
# A tibble: 1 x 3
  state evtype            max.net.loss
  <fct> <fct>                    <dbl>
1 NC    THUNDERSTORM WIND         5000

[[41]]
# A tibble: 1 x 3
  state evtype max.net.loss
  <fct> <fct>         <dbl>
1 ND    FLOOD           975

[[42]]
# A tibble: 1 x 3
  state evtype max.net.loss
  <fct> <fct>         <dbl>
1 NE    FLOOD           897

[[43]]
# A tibble: 1 x 3
  state evtype max.net.loss
  <fct> <fct>         <dbl>
1 NH    FLOOD           945

[[44]]
# A tibble: 2 x 3
  state evtype        max.net.loss
  <fct> <fct>                <dbl>
1 NJ    COASTAL FLOOD          900
2 NJ    FLOOD                  900

[[45]]
# A tibble: 1 x 3
  state evtype            max.net.loss
  <fct> <fct>                    <dbl>
1 NM    THUNDERSTORM WIND          800

[[46]]
# A tibble: 1 x 3
  state evtype max.net.loss
  <fct> <fct>         <dbl>
1 NV    FLOOD           973

[[47]]
# A tibble: 2 x 3
  state evtype    max.net.loss
  <fct> <fct>            <dbl>
1 NY    FLOOD              975
2 NY    TSTM WIND          975

[[48]]
# A tibble: 1 x 3
  state evtype    max.net.loss
  <fct> <fct>            <dbl>
1 OH    HIGH WIND          995

[[49]]
# A tibble: 1 x 3
  state evtype            max.net.loss
  <fct> <fct>                    <dbl>
1 OK    THUNDERSTORM WIND         3500

[[50]]
# A tibble: 1 x 3
  state evtype max.net.loss
  <fct> <fct>         <dbl>
1 OR    FLOOD           688

[[51]]
# A tibble: 5 x 3
  state evtype             max.net.loss
  <fct> <fct>                     <dbl>
1 PA    FLASH FLOOD                 900
2 PA    HEAVY SNOW                  900
3 PA    THUNDERSTORM WINDS          900
4 PA    TORNADO                     900
5 PA    WINTER STORM                900

[[52]]
# A tibble: 3 x 3
  state evtype             max.net.loss
  <fct> <fct>                     <dbl>
1 PH    MARINE STRONG WIND            0
2 PH    MARINE TSTM WIND              0
3 PH    WATERSPOUT                    0

[[53]]
# A tibble: 1 x 3
  state evtype           max.net.loss
  <fct> <fct>                   <dbl>
1 PK    MARINE HIGH WIND           15

[[54]]
# A tibble: 1 x 3
  state evtype     max.net.loss
  <fct> <fct>             <dbl>
1 PM    WATERSPOUT            0

[[55]]
# A tibble: 1 x 3
  state evtype    max.net.loss
  <fct> <fct>            <dbl>
1 PR    LANDSLIDE         4800

[[56]]
# A tibble: 1 x 3
  state evtype             max.net.loss
  <fct> <fct>                     <dbl>
1 PZ    MARINE STRONG WIND           25

[[57]]
# A tibble: 1 x 3
  state evtype    max.net.loss
  <fct> <fct>            <dbl>
1 RI    TSTM WIND          750

[[58]]
# A tibble: 1 x 3
  state evtype  max.net.loss
  <fct> <fct>          <dbl>
1 SC    TORNADO         1000

[[59]]
# A tibble: 1 x 3
  state evtype  max.net.loss
  <fct> <fct>          <dbl>
1 SD    TORNADO          900

[[60]]
# A tibble: 1 x 3
  state evtype           max.net.loss
  <fct> <fct>                   <dbl>
1 SL    MARINE TSTM WIND           15

[[61]]
# A tibble: 1 x 3
  state evtype       max.net.loss
  <fct> <fct>               <dbl>
1 ST    STRONG WINDS            0

[[62]]
# A tibble: 1 x 3
  state evtype  max.net.loss
  <fct> <fct>          <dbl>
1 TN    TORNADO         4410

[[63]]
# A tibble: 1 x 3
  state evtype    max.net.loss
  <fct> <fct>            <dbl>
1 TX    TSTM WIND          990

[[64]]
# A tibble: 2 x 3
  state evtype       max.net.loss
  <fct> <fct>               <dbl>
1 UT    FLASH FLOOD           900
2 UT    WINTER STORM          900

[[65]]
# A tibble: 1 x 3
  state evtype      max.net.loss
  <fct> <fct>              <dbl>
1 VA    FLASH FLOOD          900

[[66]]
# A tibble: 2 x 3
  state evtype     max.net.loss
  <fct> <fct>             <dbl>
1 VI    HEAVY RAIN          500
2 VI    HIGH WINDS          500

[[67]]
# A tibble: 1 x 3
  state evtype      max.net.loss
  <fct> <fct>              <dbl>
1 VT    FLASH FLOOD          800

[[68]]
# A tibble: 3 x 3
  state evtype           max.net.loss
  <fct> <fct>                   <dbl>
1 WA    FLOOD                     900
2 WA    LIGHTNING                 900
3 WA    WILD/FOREST FIRE          900

[[69]]
# A tibble: 1 x 3
  state evtype max.net.loss
  <fct> <fct>         <dbl>
1 WI    FLOOD           975

[[70]]
# A tibble: 1 x 3
  state evtype      max.net.loss
  <fct> <fct>              <dbl>
1 WV    FLASH FLOOD          950

[[71]]
# A tibble: 1 x 3
  state evtype max.net.loss
  <fct> <fct>         <dbl>
1 WY    HAIL            800

[[72]]
# A tibble: 1 x 3
  state evtype                   max.net.loss
  <fct> <fct>                           <dbl>
1 XX    MARINE THUNDERSTORM WIND            0

请帮我修一下。非常感谢！

Answer 1

如果希望保留顺序，可以使用levels将state列转换为factor，因为它们默认按字母顺序显示在数据中。

library(dplyr)
df$state <- factor(df$state, levels = unique(df$state))
df2 <- group_split(df)