在相同条件下为多列过滤data.table [英] Filter data.table on same condition for multiple columns

问题描述

我正在使用列名称的向量来选择data.table列的子集。我想到了是否可以在 i 中基本定义条件，然后将其应用于所有选定的列。
例如，使用 mtcars 数据集。
我想选择圆柱和齿轮柱，然后对所有具有四个圆柱和四个齿轮的汽车进行过滤。当然，我还需要为过滤器定义它是和还是或，但是我只是想知道是否这个想法可以在 data.table 上下文中应用。

I am using a vector of column names to select a subset of columns of a data.table. I had the idea if it's possible to basically define conditions in i which are then applied to all the selected columns. For example using the mtcars dataset. I would like to select the columns cylinder and gear and then would like to filter on all cars which have four cylinders and four gears. Of course I would also need to define if it is and or or for the filter, but I am just interested if the idea can be applied somehow in the data.table context.

# working code
sel.col <- c("cyl", "gear")
dt <- data.table(mtcars[1:4,])

dt[, ..sel.col]
dt[cyl == 4 & gear == 4, ..sel.col]    


# Non-working code
dt[ sel.col == 4 , ..sel.col]

推荐答案

我们可以使用 get

sel.col <- "cyl"
dt[get(sel.col) == 4, ..sel.col]
#    cyl gear
#1:   4    4

或 eval（as.name

dt[eval(as.name(sel.col)) == 4, ..sel.col]
#    cyl gear
#1:   4    4

---

以前，我们认为只有一个要评估的列，如果有多个列，请指定它在 .SDcols 中，循环遍历Data.table的子集（ .SD ），将其与感兴趣的值进行比较（'4'），通过 | Reduce 将其还原为逻辑矢量，即，每行和子集中的任何TRUE

---

Earlier, we thought that there is only a single column to be evaluated. If we have more than one column, specify it in the .SDcols, loop through the Subset of Data.table (.SD) compare it with the value of interest ('4'), Reduce it to logical vector with | i.e. any TRUE in each of the rows and subset the rows based on this

dt[dt[, Reduce(`|`, lapply(.SD, `==`, 4)),.SDcols = sel.col], ..sel.col]

这篇关于在相同条件下为多列过滤data.table的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！