r - data.table:如何根据包含列名的分组唯一行值更改列值

Question

我有一个包含 ~18^6 行的 data.table，我需要通过 ID 获取 CLASS 的唯一值，并将它们各自的列设置为 1，如下面的示例所示

DT <- data.table::data.table(ID=c("1","1","1","2","2"),
                             CLASS=c("a","a","b","c","b"),
                             a=c(0,0,0,0,0),
                             b=c(0,0,0,0,0),
                             c=c(0,0,0,0,0))

### Start with this
ID CLASS a b c
1      a 0 0 0
1      a 0 0 0
1      b 0 0 0
2      c 0 0 0
2      b 0 0 0

### Want this
ID CLASS a b c
1      a 1 1 0
1      a 1 1 0
1      b 1 1 0
2      c 0 1 1
2      b 0 1 1

我的第一直觉是尝试下面的代码，但发现它会将所有列设置为 1，因为 unique(DT$CLASS) 固有地包含所有 ID 的所有唯一值，并且不会通过“分组”参数传递说。

### Tried this
DT[,unique(DT$CLASS):=1,by=ID]

### Got this
ID CLASS a b c
1      a 1 1 1
1      a 1 1 1
1      b 1 1 1
2      c 1 1 1
2      b 1 1 1

我一直在努力充分利用 data.table 的潜力和速度，并希望仅使用 data.table 参数中的命令来创建所需的输出。

谁能帮我编写正确的代码，仅使用 data.table 命令/参数，以便我的第 j 个索引仅包含唯一值，按 ID，并将适当的列设置为 1？

Follow-up Question:

假设每一行也有一个关联的日期 RXDATE，我想创建所有类值的相应列名，这些值按 ID 按 CLASS 保存最小 RXDATE。我也可以为此求助于 dcast 吗？

### Start with this
ID CLASS a b c RXDATE
1      a 1 1 0 1-1-99
1      a 1 1 0 1-2-99
1      b 1 1 0 1-3-99
2      c 0 1 1 5-4-00
2      b 0 1 1 6-5-01

### Want this
ID CLASS a b c RXDATE   a_DT   b_DT   c_DT
1      a 1 1 0 1-1-99 1-1-99 1-3-99     NA
1      a 1 1 0 1-2-99 1-1-99 1-3-99     NA
1      b 1 1 0 1-3-99 1-1-99 1-3-99     NA
2      c 0 1 1 5-4-00     NA 6-5-01 5-4-00
2      b 0 1 1 6-5-01     NA 6-5-01 5-4-00

Answer 1

使用 dcast 和 merge 你还可以:

DT <- data.table::data.table(ID=c("1","1","1","2","2"),
                             CLASS=c("a","a","b","c","b"),
                             a=c(0,0,0,0,0),
                             b=c(0,0,0,0,0),
                             c=c(0,0,0,0,0))

# dcast to convert to wide
DT_dcast <- dcast(DT[, .(ID, CLASS)], ID ~ CLASS, fun.aggregate = function(x) length(unique(x)), value.var = "CLASS")
DT_dcast
   ID a b c
1:  1 1 1 0
2:  2 0 1 1

# Then merge with the original data.table
DT_m <- merge(DT[, .(ID, CLASS)], DT_dcast, by = "ID")
DT_m
   ID CLASS a b c
1:  1     a 1 1 0
2:  1     a 1 1 0
3:  1     b 1 1 0
4:  2     c 0 1 1
5:  2     b 0 1 1

编辑 您仍然可以对 dcast 和 merge 使用相同的方法。

我从您的“从这个开始”数据中注意到第 2 行有不同的 RX 日期，并且从“想要这个”数据中您只为此保留了“1-1-99”。

DT2 <- data.table::data.table(ID=c("1","1","1","2","2"),
                             CLASS=c("a","a","b","c","b"),
                             a=c(0,0,0,0,0),
                             b=c(0,0,0,0,0),
                             c=c(0,0,0,0,0), 
                             RXDate = c("1-1-99", "1-2-99", "1-3-99", "5-4-00", "6-5-01"))

# 2nd row from the data provided has different RXDate under same ID and Class.
# Use x[1] to pick first
DT_dcast <- dcast(DT2[, .(ID, CLASS, RXDate)], ID ~ CLASS, 
                  fun.aggregate = function(x) x[1], 
                  value.var = c("CLASS", "RXDate"))
DT_dcast
   ID CLASS.1_a CLASS.1_b CLASS.1_c RXDate_a RXDate_b RXDate_c
1:  1         a         b      <NA>   1-1-99   1-3-99     <NA>
2:  2      <NA>         b         c     <NA>   6-5-01   5-4-00

# Convert 1 or 0 under CLASS
class_cols <- names(DT_dcast)[grepl("CLASS", names(DT_dcast))]
for (col in class_cols) set(DT_dcast, j = col, value = ifelse(is.na(DT_dcast[[col]]), 0, 1))

DT_dcast
ID CLASS.1_a CLASS.1_b CLASS.1_c RXDate_a RXDate_b RXDate_c
1:  1         1         1         0   1-1-99   1-3-99     <NA>
2:  2         0         1         1     <NA>   6-5-01   5-4-00

# Then merge with the original data.table
DT_m <- merge(DT2[, .(ID, CLASS, RXDate)], DT_dcast, by = "ID")
DT_m

   ID CLASS RXDate CLASS.1_a CLASS.1_b CLASS.1_c RXDate_a RXDate_b RXDate_c
1:  1     a 1-1-99         1         1         0   1-1-99   1-3-99     <NA>
2:  1     a 1-2-99         1         1         0   1-1-99   1-3-99     <NA>
3:  1     b 1-3-99         1         1         0   1-1-99   1-3-99     <NA>
4:  2     c 5-4-00         0         1         1     <NA>   6-5-01   5-4-00
5:  2     b 6-5-01         0         1         1     <NA>   6-5-01   5-4-00

如果你想重命名列，那么你可以使用 setnames