Kaggle 데이터를 활용한 DataTable 문서화.
2019. 3. 17. 20:25ㆍ분석 R
다 돌아가지만, Posting 하기 위해서는 주석으로 올려야해서 주석으로 올립니다.
html이나 rmd가 필요하다면 댓글에 글 남겨주세요.
library(data.table)Data Loading
- fread를 사용하면 큰 데이터를 더 빠르게 읽을 수 있다.
time <- system.time(Data <- fread("./../Data/data.csv", header = TRUE))
time1 <- system.time(Data1 <- read.csv("./../Data/data.csv"))
to_table <- data.table(Data1)
to_table_2 <- setDT(Data1)
## 결측치 제거 및 Type 변환
## R에서는 잘 되지만 포스팅 할 때는 문제가 발새
#Str_to_int <- colnames(Data)[4:6]
#Data[, (Str_to_int) := lapply(.SD, as.numeric), .SDcols = Str_to_int]
Data$`Video Uploads` <- as.numeric(Data$`Video Uploads`)
Data$Subscribers <- as.numeric(Data$Subscribers)
Data$`Video views` <- as.numeric(Data$`Video views`)
colSums(is.na(Data))## Rank Grade Channel name Video Uploads Subscribers
## 0 0 0 15 344
## Video views
## 263Data <- na.omit(Data)
print(paste(class(Data) , " == " , class(Data1)))## [1] "data.table == data.table" "data.frame == data.frame"print(paste(time[3] , " < " , time1[3] ))## [1] "0 < 0.100000000000364"
Select
- Keeping Columns
Data[,1]## [1] "1st" "2nd" "3rd" "4th" "5th" "6th"
## [7] "7th" "8th" "9th" "10th" "11th" "12th"
## [13] "13th" "14th" "15th" "16th" "17th" "19th"
## [19] "20th" "21st" "22nd" "23rd" "24th" "25th"
## [25] "26th" "27th" "28th" "29th" "30th" "31st"
## [31] "32nd" "33rd" "34th" "35th" "36th" "37th"
## [37] "38th" "39th" "40th" "41st" "42nd" "43rd"
## [43] "44th" "45th" "46th" "47th" "48th" "49th"
## [49] "50th" "51st" "52nd" "53rd" "54th" "55th"
## [55] "56th" "58th" "59th" "60th" "61st" "62nd"
## [61] "63rd" "64th" "65th" "66th" "67th" "68th"
## [67] "69th" "70th" "71st" "72nd" "73rd" "74th"
Data[,1, with = F]## [1] "1st" "2nd" "3rd" "4th" "5th" "6th"
## [7] "7th" "8th" "9th" "10th" "11th" "12th"
## [13] "13th" "14th" "15th" "16th" "17th" "19th"
## [19] "20th" "21st" "22nd" "23rd" "24th" "25th"
## [25] "26th" "27th" "28th" "29th" "30th" "31st"
## [31] "32nd" "33rd" "34th" "35th" "36th" "37th"
## [37] "38th" "39th" "40th" "41st" "42nd" "43rd"
## [43] "44th" "45th" "46th" "47th" "48th" "49th"
## [49] "50th" "51st" "52nd" "53rd" "54th" "55th"
## [55] "56th" "58th" "59th" "60th" "61st" "62nd"
## [61] "63rd" "64th" "65th" "66th" "67th" "68th"
## [67] "69th" "70th" "71st" "72nd" "73rd" "74th"
## [73] "75th" "76th" "77th" "78th" "80th" "81st"
## [79] "82nd" "83rd" "84th" "85th" "86th" "87th"
## [85] "88th" "90th" "91st" "92nd" "93rd" "94th"
#Select <- Data[, Rank]
#print(head(Select))
#print(class(Select))
## multiple select
#Data[, .(Rank, Grade)]
Data[, c(1:2)]## Rank Grade
## 1: 1st A++
## 2: 2nd A++
## 3: 3rd A++
## 4: 4th A++
## 5: 5th A++
## ---
## 3783: 4,996th B+
## 3784: 4,997th B+
## 3785: 4,998th B+
## 3786: 4,999th B+
## 3787: 5,000th B+- Drop Columns
tmp2 <- Data[,with = F]
#tmp2[, !c("Rank" , "Grade")]
tmp2[, !c(1:2)]## Null data.table (0 rows and 0 cols)Order
mydata01 = Data
setorder(Data , Subscribers )
head(mydata01)## Rank Grade Channel name Video Uploads Subscribers
## 1: 4,097th B+ Xiaomi France 17 301
## 2: 93rd A hairong zheng 492 394
## 3: 3,517th B+ Jianpeng Li 232 437
## 4: 3,161st B+ Ron Funches - Topic 21 453
## 5: 2,288th B+ Christian Nodal - Topic 30 471
## 6: 3,181st B+ vendezvotrevoiture.fr 12 554
## Video views
## 1: 12943662
## 2: 254252482
## 3: 70635195
## 4: 13932203
## 5: 17868274
## 6: 17603860setorder(Data , Grade, -Subscribers )
head(mydata01)## Rank Grade Channel name Video Uploads Subscribers Video views
## 1: 175th A Dude Perfect 185 34489432 6095981611
## 2: 64th A Ed Sheeran 122 33961184 14521211040
## 3: 127th A TheEllenShow 9066 27765423 13403860667
## 4: 211th A KatyPerryVEVO 125 26712128 16244535352
## 5: 150th A EminemVEVO 80 26650488 11317532576
## 6: 84th A Felipe Neto 1276 25069826 4455604361Filter
* one condition
#head(Data[Grade == "A++"])
#head(Data[Grade == "A++",])
#head(Data[`Video views`> 50000,])- mutiple condition
# head(Data[Grade == "A++" & `Video views`> 10^10 , ])mutiple condition
Logica operator %in%
#Data[Grade %in% c("A++", "B+")]
#Data[!Grade %in% c("A++", "B+")]J Expression
기존 방식보다 더 빠르게 Filtering 가능
setkey(Data , Grade )
#Data[J("A++")]
tables()## NAME NROW NCOL MB
## 1: Data 3,787 6 1
## 2: Data1 4,393 6 1
## 3: mydata01 3,787 6 1
## 4: tmp2 3,787 6 1
## 5: to_table 4,393 6 1
## 6: to_table_2 4,393 6 1
## COLS KEY
## 1: Rank,Grade,Channel name,Video Uploads,Subscribers,Video views Grade
## 2: Rank,Grade,Channel.name,Video.Uploads,Subscribers,Video.views
## 3: Rank,Grade,Channel name,Video Uploads,Subscribers,Video views Grade
## 4: Rank,Grade,Channel name,Video Uploads,Subscribers,Video views
## 5: Rank,Grade,Channel.name,Video.Uploads,Subscribers,Video.views
## 6: Rank,Grade,Channel.name,Video.Uploads,Subscribers,Video.views
## Total: 6MBsetkey(Data , Grade , Rank)
tables()## NAME NROW NCOL MB
## 1: Data 3,787 6 1
## 2: Data1 4,393 6 1
## 3: mydata01 3,787 6 1
## 4: tmp2 3,787 6 1
## 5: to_table 4,393 6 1
## 6: to_table_2 4,393 6 1
## COLS
## 1: Rank,Grade,Channel name,Video Uploads,Subscribers,Video views
## 2: Rank,Grade,Channel.name,Video.Uploads,Subscribers,Video.views
## 3: Rank,Grade,Channel name,Video Uploads,Subscribers,Video views
## 4: Rank,Grade,Channel name,Video Uploads,Subscribers,Video views
## 5: Rank,Grade,Channel.name,Video.Uploads,Subscribers,Video.views
## 6: Rank,Grade,Channel.name,Video.Uploads,Subscribers,Video.views
## KEY
## 1: Grade,Rank
## 2:
## 3: Grade,Rank
## 4:
## 5:
## 6:
## Total: 6MB#Data[J("A++", "1st")]
#Data[J("A++")]- Condition & Summarise
# Data[Grade == "A++", mean(`Video Uploads`)]
# Data[J("A++") , mean(`Video Uploads`)]Summarise
- .N : Count
- D[,][,] Pipe 연산자와 같은 효과
# Data[, .(Rank, Grade)][,.N, by=Grade]
# Data[, tmp := "NULL"][, tmp := NULL].SD , .SDcols
- .SD : Subset of Data
# Data[, .(mean(`Subscribers`), mean(`Video views`))]
# Data[, lapply(.SD, mean)]
# Data[, lapply(.SD, mean), .SDcols = c("Subscribers","Video views")]Groupby
#names(Data)
#Data[ , mean(`Video Uploads`) , by ="Grade"]
#Data[ , max(`Video Uploads`) , by ="Grade"]
#Data[ , list(min(`Video views`) , max(`Video views`)) , by ="Grade"]Mutate
- D[ , key := func(value) , ]
## 1번째 방법
#Data[ , tmp := "Null"]
#Data[ , `:=`(tmp , "NULL")]
#Data2 <- Data[ , tmp := ifelse(`Video views` > 10^10 , "Top" , "Down")]
## 2번째 방법
#Data2[ , tmp := "Top"]
#Data2[`Video views` <= 10^10 , `:=`(tmp, "Bottom")]
#head(Data2)
#Data2[ , tmp :=NULL]
#names(Data2)Reshape
dcast
DT <- data.table(mtcars)
dcast(DT, gear ~ cyl,
value.var = c("disp", "hp"))## Aggregate function missing, defaulting to 'length'## gear disp_4 disp_6 disp_8 hp_4 hp_6 hp_8
## 1: 3 1 2 12 1 2 12
## 2: 4 8 4 0 8 4 0
## 3: 5 2 1 2 2 1 2DT2 <- dcast(DT, gear ~ cyl,
value.var = c("disp", "hp"),
fun = list(mean, sum))melt
melt(DT2,
id.vars = c("gear"),
measure.vars = patterns("^disp", "^hp"),
variable.name = "cyl",
value.name = c("disp", "hp"))## gear cyl disp hp
## 1: 3 1 120.1000 97.0000
## 2: 4 1 102.6250 76.0000
## 3: 5 1 107.7000 102.0000
## 4: 3 2 241.5000 107.5000
## 5: 4 2 163.8000 116.5000
## 6: 5 2 145.0000 175.0000
## 7: 3 3 357.6167 194.1667
## 8: 4 3 NaN NaN
## 9: 5 3 326.0000 299.5000
## 10: 3 4 120.1000 97.0000
## 11: 4 4 821.0000 608.0000
## 12: 5 4 215.4000 204.0000
## 13: 3 5 483.0000 215.0000
## 14: 4 5 655.2000 466.0000
## 15: 5 5 145.0000 175.0000
## 16: 3 6 4291.4000 2330.0000
## 17: 4 6 0.0000 0.0000
## 18: 5 6 652.0000 599.0000Merging / Joins
(dt1 <- data.table(A = letters[rep(1:3, 2)], X = 1:6, key = "A"))## A X
## 1: a 1
## 2: a 4
## 3: b 2
## 4: b 5
## 5: c 3
## 6: c 6(dt2 <- data.table(A = letters[rep(2:4, 2)], Y = 6:1, key = "A"))## A Y
## 1: b 6
## 2: b 3
## 3: c 5
## 4: c 2
## 5: d 4
## 6: d 1- inner join
merge(dt1, dt2, by="A")## A X Y
## 1: b 2 6
## 2: b 2 3
## 3: b 5 6
## 4: b 5 3
## 5: c 3 5
## 6: c 3 2
## 7: c 6 5
## 8: c 6 2- left join
merge(dt1, dt2, by="A", all.x = TRUE)## A X Y
## 1: a 1 NA
## 2: a 4 NA
## 3: b 2 6
## 4: b 2 3
## 5: b 5 6
## 6: b 5 3
## 7: c 3 5
## 8: c 3 2
## 9: c 6 5
## 10: c 6 2- Right Join
merge(dt1, dt2, by="A", all.y = TRUE)## A X Y
## 1: b 2 6
## 2: b 2 3
## 3: b 5 6
## 4: b 5 3
## 5: c 3 5
## 6: c 3 2
## 7: c 6 5
## 8: c 6 2
## 9: d NA 4
## 10: d NA 1- Full Join
merge(dt1, dt2, all=TRUE)## A X Y
## 1: a 1 NA
## 2: a 4 NA
## 3: b 2 6
## 4: b 2 3
## 5: b 5 6
## 6: b 5 3
## 7: c 3 5
## 8: c 3 2
## 9: c 6 5
## 10: c 6 2
## 11: d NA 4
## 12: d NA 1Rename
colnames(tmp2)## [1] "Rank" "Grade" "Channel name" "Video Uploads"
## [5] "Subscribers" "Video views"setnames(tmp2, c("Rank"), c("랭크"))
colnames(tmp2)## [1] "랭크" "Grade" "Channel name" "Video Uploads"
## [5] "Subscribers" "Video views"Example
## 2개 변수 같이 고려하기
#Data2[,mean_upload := mean(`Video Uploads`)]
#myfun2 = function(arg1,arg2) {
# temp1 <- ifelse(arg1 >= arg2, "평균 이상", "평균 미만")
# temp2 <- ifelse(arg1 > arg2, arg1 - arg2, arg1 / arg2)
# list(temp1,temp2)
#}
#Data2 <- Data2[, c(paste0("tmp", 4:5)) := myfun2(`Video Uploads` , mean_upload)]
#head(Data2)
#Data2[Grade == "B+", mean(Subscribers), by= "tmp4" ]
#Data2[Grade == "B+", .N , by= "tmp4"]
#Data2[Grade == "A", mean(Subscribers), by= "tmp4"]
#Data2[Grade == "A", .N, by= "tmp4"]
#Data2[,c(paste0("tmp", 4:5)) := list(NULL , NULL)]
#Data2[Grade == "A", .N]
#Data2[, length(which(Grade == "A"))]728x90
'분석 R' 카테고리의 다른 글
| [R] magick package 설치 에러 (ubuntu) (0) | 2020.05.26 |
|---|---|
| R 최신 버전 설치 관련 자료 (0) | 2020.04.29 |
| [ R ] roc curve 패키지 비교 (0) | 2019.05.01 |
| Tidyverse (dplyr , tidyr) (0) | 2019.03.17 |
| 알고리즘 체인과 파이프라인 (0) | 2018.01.25 |