帶data.table的一個(gè)組的子集假設(shè)我有一個(gè)包含一些棒球運(yùn)動(dòng)員的數(shù)據(jù)表:library(plyr)library(data.table)bdt <- as.data.table(baseball)對(duì)于每個(gè)玩家(由id給出),我希望找到與他們玩游戲最多的年份相對(duì)應(yīng)的行。這在plyr中很簡(jiǎn)單:ddply(baseball, "id", subset, g == max(g))data.table的等效代碼是什么?我試過(guò):setkey(bdt, "id") bdt[g == max(g)] # only one rowbdt[g == max(g), by = id]
# Error: 'by' or 'keyby' is supplied but not jbdt[, .SD[g == max(g)]] # only one row這樣做是可行的:bdt[, .SD[g == max(g)], by = id]但它只比plyr快30%,說(shuō)明它可能不是慣用的。
1 回答

素胚勾勒不出你
TA貢獻(xiàn)1827條經(jīng)驗(yàn) 獲得超9個(gè)贊
data.table
bdt[bdt[, .I[g == max(g)], by = id]$V1]
.SD
編輯:.SD
[.data.table
by
- 1 回答
- 0 關(guān)注
- 435 瀏覽
添加回答
舉報(bào)
0/150
提交
取消