[英]Reshaping a data in R
我有一個數據集,如文章底部所示。 數據具有四個列,分別稱為SIC,AT95Group,AT95Mean,AT95Med。 AT95Group列采用四個值,例如“ 00”,“ 01”,“ 11”和“ 10”。 當前,對於每個SIC,AT95Group的每個值都有四行。 我想以某種方式重塑數據框,以便每個SIC僅具有一行。 之前,我們為每對(SIC,AT95Group)有兩列分別稱為均值和med的列,我們想要創建本質上為4 * 2的列(對於組“ 00”,“ 11”,“ 01”,“ 10”)為4列(“均值”和“ Med”)。 八列類似於“ 00Mean”,“ 11Mean”,“ 00Med”,“ 11Med”等,每個SIC都有相應的值。
我覺得這很難做到。 有任何建議。 謝謝。
> dput(head(pp,20))
structure(list(SIC = c(1L, 1L, 1L, 10L, 10L, 10L, 10L, 12L, 12L,
12L, 12L, 13L, 13L, 13L, 13L, 14L, 14L, 14L, 14L, 15L), AT95Group = c("11",
"10", "00", "11", "01", "00", "10", "01", "11", "10", "00", "11",
"01", "00", "10", "11", "01", "10", "00", "01"), AT95Med = c(0.0691039117115276,
0.0608649722972575, 0.0609974198491522, 0.215571816296268, 0.305308985848382,
0.351312558091798, 0.352704719896703, 0.0459887720804718, 0.0304466181779069,
0.0513875431555943, 0.0541431932578377, 0.0650920855876547, 0.143724642017362,
0.156092793582293, 0.0976059673595903, 0.0116620184564053, 0.0188895210677074,
0.0356836223212195, 0.0513040852859517, 0.0982448708035204),
AT95Mean = c(0.0691039117115276, 0.0608649722972575, 0.0609974198491522,
0.215571816296268, 0.305308985848382, 0.351312558091798,
0.352704719896703, 0.0459887720804718, 0.0304466181779069,
0.0513875431555943, 0.0541431932578377, 0.0650920855876547,
0.143724642017362, 0.156092793582293, 0.0976059673595903,
0.0116620184564053, 0.0188895210677074, 0.0356836223212195,
0.0513040852859517, 0.0982448708035204)), .Names = c("SIC",
"AT95Group", "AT95Med", "AT95Mean"), row.names = c(241L, 236L,
27L, 1126L, 1035L, 1030L, 664L, 1269L, 1259L, 1245L, 1244L, 3919L,
4722L, 3329L, 3222L, 4886L, 4889L, 4951L, 4860L, 5108L), class = "data.frame")
嘗試提到的代碼的粗略嘗試失敗。 不知道如何繼續。
pp <- unique(dacc1[,c("SIC","AT95Group","AT95Med","AT95Mean")])
xsic <- unique(pp[,"SIC"]);
xlist <- list(xsic,rep("AT95",length(xsic)));
編輯:
我在運行troy的結果后獲得的結果:
> pp1 <- head(pp,20)
SIC AT95Group AT95Med AT95Mean
241 1 11 0.06910391 0.06910391
236 1 10 0.06086497 0.06086497
27 1 00 0.06099742 0.06099742
1126 10 11 0.21557182 0.21557182
1035 10 01 0.30530899 0.30530899
1030 10 00 0.35131256 0.35131256
664 10 10 0.35270472 0.35270472
1269 12 01 0.04598877 0.04598877
1259 12 11 0.03044662 0.03044662
1245 12 10 0.05138754 0.05138754
1244 12 00 0.05414319 0.05414319
3919 13 11 0.06509209 0.06509209
4722 13 01 0.14372464 0.14372464
3329 13 00 0.15609279 0.15609279
3222 13 10 0.09760597 0.09760597
4886 14 11 0.01166202 0.01166202
4889 14 01 0.01888952 0.01888952
4951 14 10 0.03568362 0.03568362
4860 14 00 0.05130409 0.05130409
5108 15 01 0.09824487 0.09824487
> molten<-melt(pp);
Using AT95Group as id variables
molten$variable<-paste(gsub("[AT95]","",molten$variable),molten$AT95Group," ");
cast(molten[,c(1,3,4)], SIC ~ variable);
> cast(molten[,c(1,3,4)], SIC ~ variable);
Error in `[.data.frame`(molten, , c(1, 3, 4)) :
undefined columns selected
我希望這個解決方案不要太神秘:
xsic <- unique(pp[,"SIC"]);
AT = c("00", "01", "10", "11")
d = data.frame(xsic=xsic);
for(i in 1:4) {
subgroup = pp[ pp$AT95Group==AT[i],];
d[[paste0(AT[i],"AT95Med")]] = subgroup$AT95Med[match(xsic,subgroup$SIC)];
d[[paste0(AT[i],"AT95Mean")]] = subgroup$AT95Mean[match(xsic,subgroup$SIC)];
}
結果:
xsic 00AT95Med 00AT95Mean 01AT95Med 01AT95Mean 10AT95Med 10AT95Mean 11AT95Med 11AT95Mean
1 0.06099742 0.06099742 NA NA 0.06086497 0.06086497 0.06910391 0.06910391
10 0.35131256 0.35131256 0.30530899 0.30530899 0.35270472 0.35270472 0.21557182 0.21557182
12 0.05414319 0.05414319 0.04598877 0.04598877 0.05138754 0.05138754 0.03044662 0.03044662
13 0.15609279 0.15609279 0.14372464 0.14372464 0.09760597 0.09760597 0.06509209 0.06509209
14 0.05130409 0.05130409 0.01888952 0.01888952 0.03568362 0.03568362 0.01166202 0.01166202
15 NA NA 0.09824487 0.09824487 NA NA NA NA
或者,您可以使用“重塑”包:
install.packages("reshape") # only run this once if you don't have it
require(reshape)
pp # this is what I called your table
molten<-melt(pp) # this stretches the table out into variable/value pairs
# then modify the "variable" values so they reflect the group (and delete 'AT95')
molten$variable<-paste(gsub("[AT95]","",molten$variable),molten$AT95Group," ")
# then use cast (you can look up the documentation in ?reshape)
# but basically this gives you a crosstab of the SICs against the new variables
# the significant of 1,3,4 is it pulls out only the columns I want to cast
cast(molten[,c(1,3,4)], SIC ~ variable)
這給你:
SIC Mean 00 Mean 01 Mean 10 Mean 11 Med 00 Med 01 Med 10 Med 11
1 1 0.06099742 NA 0.06086497 0.06910391 0.06099742 NA 0.06086497 0.06910391
2 10 0.35131256 0.30530899 0.35270472 0.21557182 0.35131256 0.30530899 0.35270472 0.21557182
3 12 0.05414319 0.04598877 0.05138754 0.03044662 0.05414319 0.04598877 0.05138754 0.03044662
4 13 0.15609279 0.14372464 0.09760597 0.06509209 0.15609279 0.14372464 0.09760597 0.06509209
5 14 0.05130409 0.01888952 0.03568362 0.01166202 0.05130409 0.01888952 0.03568362 0.01166202
6 15 NA 0.09824487 NA NA NA 0.09824487 NA NA
為了記錄,在base
還有一個reshape
函數(嗯, stats
):
reshape(pp, direction = "wide", idvar = "SIC",
timevar = "AT95Group", v.names = c("AT95Med", "AT95Mean"))
# SIC AT95Med.11 AT95Mean.11 AT95Med.10 AT95Mean.10 AT95Med.00 AT95Mean.00 AT95Med.01 AT95Mean.01
#241 1 0.06910391 0.06910391 0.06086497 0.06086497 0.06099742 0.06099742 NA NA
#1126 10 0.21557182 0.21557182 0.35270472 0.35270472 0.35131256 0.35131256 0.30530899 0.30530899
#1269 12 0.03044662 0.03044662 0.05138754 0.05138754 0.05414319 0.05414319 0.04598877 0.04598877
#3919 13 0.06509209 0.06509209 0.09760597 0.09760597 0.15609279 0.15609279 0.14372464 0.14372464
#4886 14 0.01166202 0.01166202 0.03568362 0.03568362 0.05130409 0.05130409 0.01888952 0.01888952
#5108 15 NA NA NA NA NA NA 0.09824487 0.09824487
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.